Inbreeding and sex-biased gene flow in the ant Formica exsecta

Abstract.— The objective of this study was to assess breeding and dispersal patterns of both males and females in a monogyne (a single queen per colony) population of ants. Monogyny is commonly associated with extensive nuptial flights, presumably leading to considerable gene flow over large areas. Opposite to these expectations we found evidence of both inbreeding and sex-biased gene flow in a monogyne population of Formica exsecta . We found a significant degree of population subdivision at a local scale (within islands) for queens (females heading established colonies) and workers, but not for colony fathers (the males mated to the colony queens). However, we found little evidence of population subdivision at a larger scale (among islands). More conclusive support for sex-biased gene flow comes from the analysis of isolation by distance on the largest island, and from assignment tests revealing differences in female and male philopatry. The genetic similarity between pairs of queens decreased significantly when geographical distance increased, demonstrating limited dispersal and isolation by distance in queens. By contrast, we found no such pattern for colony fathers. Furthermore, a significantly greater fraction of colony queens were assigned as having originated from the population of residence, as compared to colony fathers. Inbreeding coefficients were significantly positive for workers, but not for mother queens. The queen-male relatedness coefficient of 0.23 (regression relatedness) indicates that mating occurs between fairly close relatives. These results suggest that some monogyne species of ants have complex dispersal and mating systems that can result in genetic isolation by distance over small geographical scales. More generally, this study also highlights the importance of identifying the relevant scale in analyses of population structure and dispersal.     

Egg Cannibalism in a Facultative Polygynous Ant: Conflict for Reproduction or Strategy to Survive?

Policing, i.e. all behaviours that prevent a nestmate from reproducing, is currently observed in social insects. It is presumed to have evolved to regulate potential conflicts generated by genetic asymmetries or to enhance colony efficiency by avoiding surplus reproductives and brood. In the ant, Ectatomma tuberculatum, individual queen fecundity was similar in monogynous and polygynous colonies issued from a Mexican population. Egg cannibalism, however, occurred in the polygynous colonies. The stealing and destruction of reproductive queen‐laid eggs involved only nestmate queens, even if they were highly related. No queen appeared to monopolize reproduction in the polygynous colonies. But, the observed value of relatedness among workers differed from the expected value, suggesting an unequal sharing of reproduction between queens. We discussed whether the cannibalism of queen‐laid eggs in E. tuberculatum results from a competition for reproduction among queens or if this phenomenon is related to constraints on nutritional resources.     

Ant workers selfishly bias sex ratios by manipulating female development.

Kin selection theory predicts that social insects should perform selfish manipulations as a function of colony genetic structure. We describe a novel mechanism by which this occurs. First, we use microsatellite analyses to show that, in a population of the ant Leptothorax acervorum, workers' relatedness asymmetry (ratio of relatedness to females and relatedness to males) is significantly higher in monogynous (single-queen) colonies than in polygynous (multiple-queen) colonies. Workers rear mainly queens in monogynous colonies and males in polygynous colonies. Therefore, split sex ratios in this population are correlated with workers' relatedness asymmetry. Together with significant female bias in the population numerical and investment sex ratios, this finding strongly supports kin-selection theory. Second, by determining the primary sex ratio using microsatellite markers to sex eggs, we show that the ratio of male to female eggs is the same in both monogynous and polygynous colonies and equals the overall ratio of haploids (males) to diploids (queens and workers) among adults. In contrast to workers of species with selective destruction of male brood, L. acervorum workers therefore rear eggs randomly with respect to sex and must achieve their favoured sex ratios by selectively biasing the final caste (queen or worker) of developing females.     

Facultative polygyny in the plant-ant Petalomyrmex phylax (Hymenoptera: Formicinae): sociogenetic and ecological determinants of queen number

In polygynous ants it has been proposed that the coexistence of several queens in a colony evolved as a response to ecological, social and genetic parameters. We present demographic, histological and genetic data showing that the plant-ant Petalomyrmex phylax is facultatively and secondarily polygynous. Polygyny is functional, lowers the reproductive output per queen, and is a kin-selected trait as new queens accepted in polygynous colonies are highly related females that never left their natal colony. The degree of polygyny varies according to a geographical gradient. Northern colonies can be strongly polygynous, while at the southern edge of the species' distribution, colonies are almost exclusively monogynous. However, ecological studies of the host-plant populations revealed that this cline could not be explained by variations in the degree of nest site limitation. We discuss selective costs and benefits associated with these social structures, and propose that this cline may result from historical processes such as selection of a more dispersive strategy along a colonization front.  © 2005 The Linnean Society of London, Biological Journal of the Linnean Society , 2005, 86 , 133–151.     

Habitat age, breeding system and kinship in the ant Formica fusca

In polygyne ants (multiple queens per colony) factors that affect the distribution and survival of queens may play a key role in shaping the population-wide mating system and colony kin structure. The aim of this paper was to study the breeding system in two populations of different age in the facultatively polygyne ant Formica fusca. Both the observed numbers of queens, and the relatedness patterns among queens, workers and colony fathers were compared in two adjacent populations (ages 17 years and > 100 years) in Southern Finland. The results showed that both the mating system and colony kin structure differed between the study populations. In the old population the relatedness among workers, queens and colony fathers was high. The queens were also related to their mates, resulting in significant inbreeding in workers, but not in queens. Finally, the number of queens per colony fluctuated between years, suggesting queen turnover, and nest-mate queens shared their reproduction unequally (reproductive skew). In the younger population relatedness among queens and workers was lower than in the old population, and the colony fathers were unrelated. Furthermore, inbreeding was absent, and no conclusive evidence was found for reproductive skew among nest-mate queens. Finally, the number of queens per colony appeared more stable between years, although queen turnover occurred also in this population. The observed differences in dispersal and mating behaviour are discussed in the light of a potential connection between population age and habitat saturation.     

REPRODUCTIVE SKEW AND SPLIT SEX RATIOS IN SOCIAL HYMENOPTERA

Abstract I present a model demonstrating that, in social Hymenoptera, split sex allocation can influence the evolution of reproductive partitioning (skew). In a facultatively polygynous population (with one to several queens per colony), workers vary in their relative relatedness to females (relatedness asymmetry). Split sex‐ratio theory predicts that workers in monogynous (single‐queen) colonies should concentrate on female production, as their relatedness asymmetry is relatively high, whereas workers in the polygynous colonies should concentrate on male production, as their relatedness asymmetry is relatively low. By contrast, queens in all colonies value males more highly per capita than they value females, because the worker‐controlled population sex ratio is too female‐biased from the queens' standpoint. Consider a polygynous colony in a facultatively polygynous population of perennial, social Hymenoptera with split sex ratios. A mutant queen achieving reproductive monopoly would gain from increasing her share of offspring but, because the workers would assess her colony as monogynous, would lose from the workers rearing a greater proportion of less‐valuable females from the colony's brood. This sets an upper limit on skew. Therefore, in social Hymenoptera, skew evolution is potentially affected by queen‐worker conflict over sex allocation.     

Split sex ratios in perennial social Hymenoptera: a mixed evolutionary stable strategy from the queens' perspective?

In social Hymenoptera, relatedness asymmetries due to haplodiploidy often generate conflicts of genetic interest between queens and workers. Split sex ratios are common in ant populations and may result from such conflicts, with workers favoring the production of males in some colonies and of gynes in others. Such intercolonial differences may result from variations in relatedness asymmetries among colony members, but several examples are now known in which this hypothesis does not hold. We develop here a simple model assuming monogynous, monoandrous, worker-sterile, perennial colonies without dispersal restrictions. Workers may eliminate eggs of either sex and determine the caste of the female brood, but the queen controls the number of eggs of each sex she lays. In such conditions, we demonstrate that split sex ratios can result from queens adopting a mixed evolutionary stable strategy (ESS), with one option being to put a strict limit to the number of diploid eggs available and the alternative one to provide diploid eggs ad lib. In the former situation, workers should raise all diploid eggs as workers and release only male sexuals. In the latter, workers should adjust the caste ratio so as to reach the maximum sexual productivity for the colony, which is entirely invested into gynes. For a particular relative investment in gynes at the population level, between 0.5 (ESS under full queen control) and 0.75 (ESS under full worker control), an equilibrium is reached at which both strategies yield an equal genetic payoff to the queen. Male-specialized colonies are predicted to be equally abundant but less populous and less productive than gyne-specialized ones. Available data on the monogyne form of the fire ant, Solenopsis invicta, suggest that this model may apply in this case, although more specific studies are required to test these predictions.     

Facultative polygyny in Ectatomma tuberculatum (Formicidae, Ectatomminae)

Summary. Polygyny, the presence of several mated queens within the same colony, is widespread in insect societies. This phenomenon is commonly associated with ecological constraints such as limited nest sites. In habitats where solitary nest foundation is risky, monogynous colonies can reintegrate young daughter queens (secondary polygyny). We studied the reproductive structure (i.e. queen number) of the ectatommine ant Ectatomma tuberculatum from Bahia State, Brazil. This species was found to present facultative polygyny: out of a total of 130 colonies collected, 39.2% were monogynous, while 43.8% were polygynous. Polygynous colonies had significantly more workers than monogynous ones. Queen number in polygynous colonies ranged from 2 to 26, with an average of 4 ± 4 queens per colony. All nestmate queens were egg-layers with no apparent dominance hierarchy or agonistic behavior. Non-nestmate queens were adopted by monogynous colonies suggesting that polygyny is secondary, originating through queen adoption. This species is characterized by an open recognition system, which probably allows a switch from monogynous to polygynous colonies. The behavioral acts of queens showed that resident queens remained frequently immobile on or near the brood, contrarily to alien or adopted queens and gynes. In addition, monogynous queens showed no behavioral or physiological (i.e. by ovarian status) differences in comparison with polygynous ones. Secondary or facultative polygyny, probably associated with queen adoption, may have been favored in particular environmental conditions. Indeed, by increasing colony productivity (i.e. number of workers) and territory size (by budding and polydomy), polygyny could uphold E. tuberculatum as a dominant species in the mosaic of arboreal ants in Neotropical habitats.Received 7 April 2004; revised 11 November 2004; accepted 15 November 2004.     

Hierarchical analysis of population genetic structure in the monogynous ant Cataglyphis cursor using microsatellite and mitochondrial DNA markers

Despite having winged queens, female dispersal in the monogynous ant Cataglyphis cursor is likely to be restricted because colonies reproduce by fission. We investigated the pattern of population genetic structure of this species using eight microsatellite markers and a mitochondrial DNA (mtDNA) sequence, in order to examine the extent of female and nuclear gene flow in two types of habitat. Sampling was carried out at a large spatial scale (16 sites from 2.5 to 120 km apart) as well as at a fine spatial scale (two 4.5-km transects, one in each habitat type). The strong spatial clustering of mtDNA observed at the fine spatial scale strongly supported a restricted effective female dispersal. In agreement, patterns of the mtDNA haplotypes observed at large and fine spatial scales suggested that new sites are colonized by nearby sites. Isolation by distance and significant nuclear genetic structure have been detected at all the spatial scales investigated. The level of local genetic differentiation for mitochondrial marker was 15 times higher than for the nuclear markers, suggesting differences in dispersal pattern between the two sexes. However, male gene flow was not sufficient to prevent significant nuclear genetic differentiation even at short distances (500 m). Isolation-by-distance patterns differed between the two habitat types, with a linear decrease of genetic similarities with distance observed only in the more continuous of the two habitats. Finally, despite these low dispersal capacities and the potential use of parthenogenesis to produce new queens, no signs of reduction of nuclear genetic diversity was detected in C. cursor populations.     

Variation in patriline reproductive success during queen production in orphaned colonies of the thelytokous ant Cataglyphis cursor

In genetically diverse insect societies (polygynous or polyandrous queens), the production of new queens can set the ground for competition among lineages. This competition can be very intense when workers can reproduce using thelytoky as worker lineages that manage to produce new queens gain a huge benefit. Selection at the individual level might then lead to the evolution of cheating genotypes, i.e. genotypes that reproduce more than their fair share. We studied the variation in reproductive success among worker patrilines in the thelytokous and highly polyandrous ant Cataglyphis cursor. Workers produce new queens by thelytoky in orphaned colonies. The reproductive success of each patriline was assessed in 13 orphaned colonies using genetic analysis of 433 workers and 326 worker-produced queens. Our results show that patrilines contributed unequally to queen production in half of the colonies, and the success of patrilines was function of their frequencies in workers. However, over all colonies, we observed a significant difference in the distribution of patrilines between workers and worker-produced queens, and this difference was significant in three of 13 colonies. In addition, six colonies contained a low percentage of foreign workers (drifters), and in one colony, they produced a disproportionably high number of queens. Hence, we found some evidence for the occurrence of rare cheating genotypes. Nevertheless, cheating appears to be less pronounced than in the Cape Honey bee, a species with a similar reproductive system. We argue that worker reproduction by parthenogenesis might not be common in natural populations of C. cursor.     

Pirate ants (Polyergus breviceps) and sympatric hosts (Formica occulta and Formica sp. cf. argentea): host specificity and coevolutionary dynamics

The pace and trajectory of coevolutionary arms races between parasites and their hosts are strongly influenced by the number of interacting species. In environments where a parasite has access to more than one host species, the parasite population may become divided in preference for a particular host. In the present study, we show that individual colonies of the pirate ant Polyergus breviceps differ in host preference during raiding, with each colony specializing on only one of two available Formica host species. Moreover, through genetic analyses, we show that the two hosts differ in their colony genetic structure. Formica occulta colonies were monogynous, whereas Formica  sp. cf. argentea colonies were polygynous and polydomous (colonies occupy multiple nest sites). This difference has important implications for coevolutionary dynamics in this system because raids against individual nests of polydomous colonies have less impact on overall host colony fitness than do attacks on intact colonies. We also used primers that we designed for four microsatellite loci isolated from P. breviceps to verify that colonies of this species, like other pirate ants, are comprised of simple families headed by one singly mated queen.  © 2007 The Linnean Society of London, Biological Journal of the Linnean Society , 2007, 91 , 565–572.     

Dynamics and Consequences of Hierarchy Formation in the Ant Leptothorax sp. A

In the Nearctic ant Leptothorax sp. A, aggressive interactions among wingless intermorphic queens and primarily winged gynomorphic queens lead to the formation of dominance hierarchies, in which the highest-ranking individual is the only egg-layer in a colony. Fighting occurs during two periods of the annual cycle: in late summer, newly adopted young queens are integrated into the colony's hierarchy; after hibernation, fighting resumes and the high aggressiveness of α-queens may now lead to the emigration of β and other middle-ranking queens. The α-position appears to be very stable over successive fighting periods, though an estimate of nestmate relatedness by allozyme electrophoresis (Polyacrylamide gels and cellulose acetate plates) suggests that queen replacement occasionally occurs. The mean relatedness determined for adult workers in a functionally monogynous population of Leptothorax sp. A was 0.54 and their effective number of mothers therefore 1.5. This is lower than expected and found for monogynous colonies. Dominance rank is apparently not correlated with queen morph, weight, and size, but an influence of insemination, age, or previous reproductive experience is likely.     

Sex ratios and the distribution of elaiosomes in colonies of the ant, Aphaenogaster rudis

Summary: Genetic theory predicts that workers in monogynous ant colonies with singly-mated queens should capitalize on higher relatedness with sisters than with brothers by altering the sex investment ratio of a colony in favor of females. Sex investment ratios, however, may also be influenced by the amount of resources available to colonies, in part because more mating opportunities might be obtained by investing scarce resources in males, which are much smaller than queens. Female larvae that reach a critical size by a particular point in development become queens while underfed larvae develop into workers, so workers could potentially influence the sex investment ratio of a colony by selectively feeding female larvae. In a previous experiment on the ant, Aphaenogaster rudis, colonies increased female sex investment after their diet was supplemented with elaiosomes, a lipid-rich food gained from a seed dispersal mutualism. In order to investigate the mechanisms producing this shift, we radio-labeled Sanguinaria canadensis elaiosomes with fatty acids and compared uptake among castes within a colony. The experiment was performed in both the laboratory and field. Lab colonies produced female-biased sex investment ratios, while field colonies mainly invested in males. We hypothesize that this discrepancy is related to differing levels of background food availability in the lab and field. The results of the elaiosome distribution experiment do not support a hypothesis that elaiosomes play a qualitative role in queen determination, because all individuals in a colony receive this nutrient. There is, however, support for the hypothesis that elaiosomes have a quantitative effect on larval development because larvae that accumulated more radio-label from elaiosomes tended to develop into gynes (virgin queens), while other female larvae developed into workers.     

Nutritional versus genetic correlates of caste differentiation in a desert ant

1. In many ant species, caste differentiation stems from trophic differences at the larval stage. Adult workers that feed larvae have great control over the allocation of colony resources to growth (production of workers) versus reproduction (production of queens). However, larval caste fate may also be constrained very early on through direct genetic effects or non‐genetic maternal effects. 2. Here, we combined isotopic and genetic analyses to study the developmental origin of queens and workers in a desert‐dwelling ant, Cataglyphis tartessica (Amor & Ortega, 2014). Queens do not found new colonies alone but rather disperse with workers. As the latter are always wingless, selection pressures on specific queen traits such as flight ability have become relaxed. Though the phylogenetically related species, C. emmae (Forel, 1909) only produces winged queens much larger than workers, C. tartessica produces two types of small queens relative to workers: brachypterous (short‐winged) queens and permanently apterous ergatoid (wingless and worker‐like) queens. 3. Upon emergence, workers and ergatoids have similar δ15N isotopic values, which were lower than those of brachypters, suggesting the latter are fed more protein as larvae. Microsatellite analyses indicated that: (i) colonies are mostly monogynous and monandrous; (ii) both ergatoids and brachypters are equally related to workers; and (iii) in the few polyandrous colonies, patrilines were evenly represented across workers, brachypters and ergatoids. 4. Overall, there was no evidence of genetic caste determination. We suggest that, in contrast to brachypters, ergatoids are selfish individuals that escape the nutritional castration carried out by workers and develop into queens in spite of the colony's collective interests.     

Colony sex ratios vary with breeding system but not relatedness asymmetry in the facultatively polygynous ant Pheidole pallidula

Abstract.— We investigated sex allocation in a Mediterranean population of the facultatively polygynous (multiple queen per colony) ant Pheidole pallidula . This species shows a strong split sex ratio, with most colonies producing almost exclusively a single-sex brood. Our genetic (microsatellite) analyses reveal that P. pallidula has an unusual breeding system, with colonies being headed by a single or a few unrelated queens. As expected in such a breeding system, our results show no variation in relatedness asymmetry between monogynous (single queen per colony) and polygynous colonies. Nevertheless, sex allocation was tightly associated with the breeding structure, with monogynous colonies producing a male-biased brood and polygynous colonies almost only females. In addition, sex allocation was closely correlated with colony total sexual productivity. Overall, our data show that when colonies become more productive (and presumably larger) they shift from monogyny to polygyny and from male production to female production, a pattern that has never been reported in social insects.     

Population genetics of the socially polymorphic ant <Emphasis Type="Italic">Formica podzolica</Emphasis>

Summary. We used microsatellite markers to analyze the hierarchical genetic structure of the North American mound building ant, Formica podzolica. About one-third of all colonies were headed by a single queen (monogynous) whose effective mating frequency was close to one (nestmate worker relatedness r = 0.70), while the remaining colonies were polygynous, with low average nestmate relatedness (r = 0.16). The low worker relatedness found in most polygynous colonies furthermore suggested that the numbers of queens in polygynous colonies of this ant are usually high. Contrary to what has been described from other ants with a queen number dichotomy, we did not find an effect of social form variation on the partitioning of genetic variation above the level of the colony. We found no significant differentiation between the sympatric social forms of F. podzolica, nor did differentiation among populations appear to be affected by colony social organization. These unexpected patterns of genetic structure may have resulted from differences either in the spatial distribution of the social forms or in their social flexibility.Received 12 January 2004; revised 23 February 2004; accepted 10 March 2004.     

Back to one: consequences of derived monogyny in an ant with polygynous ancestors

The number of queens per colony is of fundamental importance in the life history of social insects. Multiple queening (polygyny), with dependent colony founding by budding, has repeatedly evolved from ancestral single queening (monogyny) and independent founding by solitary queens in waSPS, bees and ants. By contrast, the reversal to monogyny appears to be rare, as polygynous queens often lack morphological adaptations necessary for dispersal and independent colony founding. In the ant genus Cardiocondyla, monogynous species evolved from polygynous ancestors. Here, we show that queens of monogynous species found their colonies independently, albeit in an unusual way: they mate in the maternal nest, disperse on foot and forage during the founding phase. This reversal appears to be associated with the occurrence of a wing polymorphism, which reflects a trade-off between reproduction and dispersal. Moreover, queens of monogynous species live considerably longer than queens in related polygynous taxa, suggesting that queen life span is a plastic trait.     

Colony structure of a slavemaking ant. I. Intracolony relatedness, worker reproduction, and polydomy

Colony and population structure of the obligate slavemaker ant Protomognathus americanus was analyzed via four nuclear microsatellite loci and mitochondrial DNA (mtDNA) markers. Colonies of P. americanus usually contain a single queen, and here we show that she is singly inseminated. Nestmate workers are generally full sisters and their relatedness does not deviate from the expected value of 0.75. Even though colonies were strictly monogynous, we were able to infer that colony takeover by related queens was common and queen replacement by unrelated queens was rare. Polydomy is widespread, with neighboring nests having the same genetic composition. Although we found no evidence of population viscosity or inbreeding from nuclear markers, mtDNA markers provided evidence for small-scale genetic structuring. Haplotype structuring and takeover by related queens suggest philopatry of newly mated queens. In this species, workers reproduce in queenright and queenless nests and worker reproduction accounts for more than 70% of all males. Although sex-ratio theory points to slavemaking ants as important systems for studying queen-worker conflict, our results indicate no basis for such conflict in P. americanus, because extensive worker reproduction generates shifts in relatedness values. Rather, the dual effects of independent polydomous nest units and local resource competition among queens produce male-biased allocation ratios in this species.     

Restricted effective queen dispersal at a microgeographic scale in polygynous populations of the ant Formica exsecta

Ecological constraints on effective dispersal have been suggested to be a key factor influencing social evolution in animal societies as well as the shift from single queen colonies (monogyny) to multiple queen colonies (polygyny) in ants. However, little is known about the effective dispersal patterns of ant queens. Here we investigate the microgeographic genetic structure of mitochondrial haplotypes in polygynous populations of the ant Formica exsecta, both between pastures and among nests within pastures. An analysis of molecular variance revealed a very high genetic differentiation (phiST = 0.72) between pastures, indicating that queens rarely disperse successfully between pastures, despite the fact that pastures were sometimes as close as 1 km. Most of the pastures contained only a single haplotype, and haplotypes were frequently distinct between nearby pastures and even between groups of nests within the same pasture. In the three pastures that contained several haplotypes, haplotypes were not randomly distributed, the genetic differentiation between nests being phiST = 0.17, 0.52, and 0.69. This indicates that most queens are recruited within their parental colonies. However, a large proportion of nests contained more than one haplotype, demonstrating that colonies will sometimes accept foreign queens. The relatedness of mitochondrial genes among nestmates varied between 0.62 and 0.75 when relatedness was measured within each pasture and ranged between 0.72 and 1.0 when relatedness was assessed with all pastures as a reference population. Neighboring nests were more genetically similar than distant ones, and there was significant isolation by distance. This pattern may be due to new nests being formed by budding or by limited effective queen dispersal, probably on foot between neighboring nests. These results show that effective queen dispersal is extremely restricted even at a small geographical scale, a pattern consistent with the idea that ecological constraints are an important selective force leading to the evolution and maintenance of polygyny.     

Mating structure and male production in Vespa analis and Vespa simillima (Hymenoptera: Vespidae)

We estimated queen mating frequency, genetic relatedness between workers and worker reproduction in the hornets Vespa analis and Vespa simillima using microsatellite DNA genotyping. The 20 V. analis colonies studied each contained a queen inseminated by a single male. Of the 15 V. simillima colonies studied, nine had a queen inseminated by a single male, four had a queen inseminated by two males, and two had a queen inseminated by three males. The estimated effective number of matings was 1.33 ± 0.74 (mean ± SD), with 75–85% of the offspring of the six multiply mated queens sired by single males. The values for genetic relatedness between the workers of V. analis and V. simillima were 0.739 ± 0.004 and 0.698 ± 0.013 (mean ± SD), respectively. We conclude that V. analis and V. simillima colonies are genetically monogynous and monandrous. When high relatedness between the workers occurs within colonies, kin selection theory predicts a potential conflict between queens and workers over male production. To determine whether males were derived from queens or workers, males from V. analis and V. simillima colonies were genotyped at four microsatellite loci and the level of ovary activation in workers was determined. None of the 787 V. analis workers and only 15 of 3520 V. simillima workers had developed ovaries. Furthermore, the genotyping identified no worker‐produced males in any colony. The presence of reproductive workers correlated positively with the number of workers within the colony. These results suggest that eusocial colonies with an annual life cycle tend to break down socially when they become large and are close to dying.