• admixture;
  • Apis mellifera ;
  • breeding;
  • management


  1. Top of page
  2. Abstract
  3. Comment
  4. Acknowledgements
  5. References

De la Rúa et al. (2013) express some concerns about the conclusions of our recent study showing that management increases genetic diversity of honey bees (Apis mellifera) by promoting admixture (Harpur et al. 2012). We provide a brief review of the literature on the population genetics of A. mellifera and show that we utilized appropriate sampling methods to estimate genetic diversity in the focal populations. Our finding of higher genetic diversity in two managed A. mellifera populations on two different continents is expected to be the norm given the large number of studies documenting admixture in honey bees. Our study focused on elucidating how management affects genetic diversity in honey bees, not on how to best manage bee colonies. We do not endorse the intentional admixture of honey bee populations, and we agree with De la Rúa et al. (2013) that native honey bee subspecies should be conserved.


  1. Top of page
  2. Abstract
  3. Comment
  4. Acknowledgements
  5. References

We recently showed that managed populations of the honey bee Apis mellifera have high levels of admixture and consequently have higher levels of genetic diversity relative to their progenitors (Harpur et al. 2012). In their comment, De la Rúa et al. (2013) raise the following concerns about the generality and interpretation of our findings: (i) our sampling scheme is insufficient for making general conclusions; (ii) managed bees are not always admixed; (iii) we made arbitrary distinctions between ‘progenitor’ and ‘managed’ bees, which is difficult given extensive gene flow between the two; (iv) comparisons between African and European bees suggest that management decreases diversity; (v) managed bees don't always have reduced diversity; (vi) we ignored the relationship between within-colony diversity and colony health; (vii) we endorsed admixture as a management practice. We believe that most of De la Rúa et al. (2013)'s concerns stem from a misunderstanding of our study and previous studies of honey bee population genetics employing single nucleotide polymorphism (SNP) markers, as we explain below. We stress that admixture (i.e. mixed genetic ancestry) is the norm in managed bee populations, and this is both theoretically and empirically expected to increase genetic diversity in managed bees relative to their progenitors. We never endorsed admixture as a management tool and agree with De la Rúa et al. (2013) regarding the need to conserve the genetic heritage of native honey bee subspecies.

In arguably the most comprehensive study of the population genetics of the honey bee, Whitfield et al. (2006) genotyped 341 bees from 14 subspecies at 1136 SNPs. Whitfield et al. (2006) found that honey bees strongly cluster into four previously hypothesized evolutionary lineages in Africa (A group, including A. m. scutellata and several other subspecies), in West and North Europe (M group, including A. m. mellifera and A. m. iberiensis), in East Europe (C group, A. m. ligustica and A. m. carnica) and in Asia (O group, multiple subspecies) (Ruttner 1988). Harpur et al. (2012) used the common designations ‘East European’ for C group bees and ‘West European’ for M group bees, following Whitfield et al. (2006) and others. We realize that this terminology can be ambiguous because the historical distribution of A. m. mellifera is broad, and we sampled this subspecies from Poland; here, we strictly refer to C and M group bees. Whitfield et al. (2006) observed very high levels of genetic differentiation between groups and low levels of genetic differentiation between subspecies/populations within groups (Whitfield et al. 2006; Zayed & Whitfield 2008), and this finding was independently replicated by our group (Kent et al. 2011; Harpur et al. 2012; Harpur & Zayed 2013). For example, Whitfield et al. (2006) did not observe any fixed differences between subspecies within the C lineage (A. m. ligustica and A. mcarnica) and M lineage (Ammellifera and A. m. iberiensis) at any of the 1136 SNPs assayed. De la Rúa et al. (2013) criticize us for not sampling populations from different localities in Europe and Africa (Comment 1) and suggested that our sampling is insufficient for making generalizations. However, De la Rúa et al. (2013)'s concerns were not realized given that most of the genetic differences in honey bees occur between the different lineages and not between subspecies within lineages or between different geographic regions within lineages (Whitfield et al. 2006). Considering that, within a lineage, bees from different subspecies and geographic regions cluster into the same ‘population’ based on unsupervised Bayesian analysis of population structure (Whitfield et al. 2006; Kent et al. 2011; Harpur et al. 2012), Harpur et al. (2012)'s sampling is appropriate for estimating genetic diversity within C, M and A group bees.

De la Rúa et al. (2013) state that ‘management of honeybees does not necessarily mean admixture, as many beekeepers around the world exclusively breed bees native to their areas’ (Comment 2). We agree that some beekeepers can and do exclusively breed and manage ‘pure’ native bees, but we believe that such practices are rare given the overwhelming population genetic evidence for admixture in honey bees. Whitfield et al. (2006) found a substantial and consistent level of admixture across managed honey bee populations in North and South America, which have high ancestries to both M and C group bees that were introduced to the new world from Europe; a similar pattern was observed by Harpur et al. (2012). Whitfield et al. (2006) also found evidence of admixture in ‘pure’ subspecies from Europe; for example, all A. m. ligustica and A. m. carnica samples had varying degrees of introgression of M group alleles; a similar pattern was observed by Harpur et al. (2012). Indeed, varying degrees of admixture/introgression were observed in most studies of the population genetics of honey bees using a variety of markers (Dall'Olio et al. 2007; De la Rua et al. 2001; reviewed by De la Rua et al. 2009; Franck et al. 2000; Garnery et al. 1998; Harpur et al. 2012; Jensen et al. 2005; Kent et al. 2011; Koulianos & Crozier 1996; Muñoz et al. 2012; Oldroyd et al. 1995; Oleksa et al. 2011; reviewed by Rortais et al. 2011; Soland-Reckeweg et al. 2009; and many more); admixture certainly would seem to be a general phenomenon in honey bees! This high degree of introgression is a byproduct of human-assisted movement of honey bees around the globe and is enhanced by the honey bee's mating biology, where queens naturally mate with a large number of unrelated drones from a large geographic area (Baudry et al. 1998).

Despite De la Rúa et al. (2013)'s assertion that many beekeepers utilize ‘pure’ native bees, they are fully aware of the extreme difficulty of maintaining ‘pure’ stocks, given ‘…extensive gene flow that occurs between managed and wild/feral honeybee populations’, ‘… the notorious problems in achieving controlled mating’ and ‘… few breeders can afford strict control measures using islands as mating stations or instrumental insemination’ (De la Rúa et al. 2013; italics ours). As such, both empirical population genetic data and knowledge of common management practices (Oldroyd 2012; De la Rúa et al. 2013) suggest that admixture is a general characteristic of managed honey bee populations. Although we are aware that ‘pure’ honey bee populations still exist in Europe, we note that the distribution of native subspecies is restricted and that such populations are at a constant threat from introgression from managed colonies (De la Rua et al. 2009), which is precisely why De la Rua et al. (2009, 2013) are concerned about the conservation of native A. mellifera subspecies in Europe. Despite De la Rúa et al. (2013)'s criticism regarding the generality of admixture in honey bees, several of the authors had previously noted that ‘Large scale migratory bee-keeping and trade in queens, coupled with the promiscuous mating system of honeybees, have exposed native European honeybees to increasing introgressive hybridization with managed non-native subspecies’ (De la Rua et al. 2009); a statement that clearly underscores the prevalence of admixture in honey bees.

Given that managed bees are mostly derived from the C and M progenitors in Europe, and because admixture is expected to be greatest in actively managed honey bee populations (i.e. because of regular stock importation, migratory beekeeping and the use of naturally mated queens), we decided to compare genetic diversity of managed honey bees from two populations in Ontario and France relative to their progenitor populations, which are now best approximated by sampling ‘pure’ M and C group bees. We obtained M and C group bees from academic researchers who had previous knowledge of the purity of local bee populations. Nevertheless, we objectively estimated the population origin and admixture level in all sampled bees using unsupervised Bayesian analyses (i.e. we did not use prior population/locality data to infer population genetic structure or admixture levels). We found higher levels of admixture and diversity in managed populations relative to mostly pure C and M group bees, leading us to conclude that management increases diversity by promoting admixture. De la Rúa et al. (2013) criticize us for detecting low admixture in bees from central Europe, but these samples were intentionally collected from ‘pure’ populations.

De la Rúa et al. (2013) also criticize us for arbitrarily classifying populations as ‘managed’ and ‘progenitor’ and for excluding progenitor bees with high levels of admixture (Comment 3). To address De La Rúa et al.'s concerns, we now present an alternative way to examine the data set. Instead of comparisons between pure ‘progenitor’ populations and ‘managed’ populations, we can instead examine the relationship between admixture and genetic diversity in European bees and their descendants in North America (Fig. 1). There is a strong and highly significant relationship between admixture and genetic diversity in European bees and their derivatives (M and C, and bees from France and Canada). Increased diversity in admixed populations is of course expected when two or more distinct populations interbreed. Studies on invasive species have demonstrated that multiple introductions from different source populations substantially elevate the genetic diversity found in invasive populations (reviewed by Dlugosch & Parker 2008). For example, the grass Phalaris arundinacea was introduced to North America from different source populations in Europe resulting in invasive populations with significantly higher levels of diversity than their progenitors in Europe (Lavergne & Molofsky 2007). Similarly, multiple introductions from different source populations led to the establishment of a invasive anole population in North America, with substantially more genetic diversity than native progenitors in the Caribbean (Kolbe et al. 2004). In the honey bee, M and C group bees show the greatest levels of genetic divergence (pairwise Fst = 0.64) (Zayed & Whitfield 2008); the two groups exhibit marked differences in allele frequencies at most loci surveyed. Beekeeping activities have resulted in substantial mixing of the two most genetically divergent honey bee populations, which results in admixed bees with higher levels of genetic diversity. Our data (Fig 1), along with previous studies showing moderate to high degrees of admixture in managed bee populations, suggest that common management practices (e.g. stock importation, migratory beekeeping and use of naturally mated queens) can serve to increase genetic diversity in managed honey bee populations by promoting admixture.


Figure 1. Admixture is significantly associated with observed heterozygosity in individual honey bees of European descent (●, bees sampled from Europe and Ontario Canada; r2 = 42.9%, p = 1.42 × 10−5). Admixture proportions were estimated using STRUCTURE (Pritchard et al. 2000) as described in Harpur et al. (2012). Observed heterozygosity was estimated as the proportion of heterozygous SNPs per bee. We previously presented this data as a contrast between ‘pure’ progenitor populations from Europe (either M or C) versus ‘admixed’ managed populations from France and Canada (a mix of M and C) (Harpur et al. 2012, fig. 2). African bees (□) are characterized by a high diversity and low admixture as shown previously.

Download figure to PowerPoint

Although De la Rúa et al. (2013) ‘agree that admixture can enhance local genetic diversity’, they argue that reduced diversity of European bees (M, C and their descendants in the new world) relative to African bees suggests that management decreases genetic diversity (Comment 4), and that several studies have found that management does not affect diversity (Comment 5). As we discuss in Harpur et al. (2012), comparisons between African and European honey bees are not appropriate because management is confounded with demography history. Honey bees colonized Europe via two independent expansions from Africa or the Middle East (Whitfield et al. 2006; Han et al. 2012), and European populations then experienced reductions in size during glacial periods (Ruttner 1988; Garnery et al. 1998). Similar to other organisms (Cornuet & Luikart 1996; Garrigan & Hammer 2006), founder events/bottlenecks associated with colonization and recolonizaton of Europe are expected to significantly reduce genetic diversity in European bees relative to African bees without having to invoke management. The authors cite a study by Dall'Olio et al. (2007) that they claim shows ‘a notorious reduction of genetic diversity in Italian honeybee populations’, but Dall'Olio et al. (2007, see table 2) actually found typical patterns of genetic diversity in Italian bees relative to other sampled populations. De la Rúa et al. (2013) also suggest that two of their previous studies (Muñoz & De la Rúa 2012; Muñoz et al. 2012) – showing that queen introductions did not increase genetic diversity of two island populations – are inconsistent with our finding that management increases diversity. However, the two studies (Muñoz & De la Rúa 2012; Muñoz et al. 2012) did not sample the source/progenitor populations used to establish the island populations; it is possible that the island populations have more diversity relative to their progenitors. The question here is not whether managed bees have low or high diversity, but rather, whether managed bees have less or more diversity than their progenitor populations.

We found high levels of genetic diversity in managed bees relative to their progenitors, which counters the idea that reduced genetic diversity is responsible for global declines of managed colonies – such colonies actually have more diversity than the populations they were derived from. De la Rúa et al. (2013) suggest that we ignore a rich body of literature showing that within-colony genetic diversity increases fitness in bees (Comment 6). However, Harpur et al. (2012) actually state that ‘Within-colony genetic diversity is clearly important …’, but we cautioned that there has been no direct evidence showing that bee population declines are associated with reduced within-colony diversity. We are aware of the effects of decreased genetic diversity at the sex determination locus in many Hymenoptera, which increases the production of inviable/sterile diploid males and affects several important population parameters (Zayed 2004, 2009; Zayed & Packer 2005; Heimpel & de Boer 2008; Harpur et al. 2013). However, declining honey bee populations do not exhibit symptoms related to diploid male production (i.e. shot brood) (vanEngelsdorp et al. 2009; Currie et al. 2010; van der Zee et al. 2012), and we did not find any evidence of inbreeding (i.e. significantly elevated Fis) in the two managed populations sampled (Harpur et al. 2012). Moreover, De la Rúa et al. (2013)'s criticism is not realized given that managed bees have more (and not less) diversity than their progenitors (Harpur et al. 2012). By eliminating decreased genetic diversity as a possible cause for colony declines, the field can now concentrate its efforts on more likely culprits (vanEngelsdorp et al. 2009; Currie et al. 2010; Cresswell et al. 2012; van der Zee et al. 2012).

Finally, De la Rúa et al. (2013) are concerned that our study justifies admixture of honey bee populations (Comment 7). It was not our intention to make recommendations about bee management practices. Our study was designed to test the long-standing hypothesis that management decreases genetic diversity in A. mellifera; we found the opposite. Our statement ‘Beekeepers may be, intentionally or unintentionally, selecting hybrid colonies, which tend to have higher fitness at some-colony-level traits’ (Harpur et al. 2012) is supported by a large number of studies documenting the benefits of within-colony genetic diversity on colony fitness (Tarpy 2003; Mattila & Seeley 2007; Oldroyd & Fewell 2007). Nevertheless, we certainly would not – and indeed did not – explicitly advocate admixture as a management tool for increasing genetic diversity. Although typical management practices do not reduce genetic diversity in honey bees, management can potentially affect the health of honey bee colonies via other mechanisms (e.g. reduced nutrition or enhanced disease transmission). As De la Rúa et al. (2013) suggest, admixture threatens to homogenize honey bee populations, which can possibly lead to the loss of local adaptations, and perhaps enhance the spread of pathogens/diseases in the resulting homogenous populations. We agree with De la Rua et al. (2009, 2013) that native honey bee subspecies should be conserved – they have proven to be excellent resources for understanding the evolution of social behaviour and worker division of labour (Kent et al. 2012; Zayed & Robinson 2012) in addition to their value for beekeeping (Cobey et al. 2012; Sheppard 2012).

In summary, Harpur et al. (2012) utilized appropriate sampling and analytical methods to compare genetic diversity in managed honey bees and their European progenitors. Harpur et al. (2012)'s argument that management promotes admixture is strongly supported by a large number of studies documenting admixture in honey bees and by knowledge of common beekeeping practices (i.e. most beekeepers/breeders do not exercise control over queen mating). Higher genetic diversity in admixed honey bee populations (Harpur et al. 2012) should not come as a surprise. Our findings are precisely what is expected when humans facilitate the movement and interbreeding of previously structured wildlife populations.


  1. Top of page
  2. Abstract
  3. Comment
  4. Acknowledgements
  5. References

BAH is funded by an Elia Scholarship from York University, while AZ is funded by a NSERC Discovery Grant and an Early Researcher Award from the Ontario Ministry of Research and Innovation. We thank two anonymous reviewers for helpful comments on the manuscript.


  1. Top of page
  2. Abstract
  3. Comment
  4. Acknowledgements
  5. References
  • Baudry E, Solignac M, Garnery L et al. (1998) Relatedness among honeybees (Apis mellifera) of a drone congregation. Proceedings of the Royal Society B-Biological Sciences, 265, 20092014.
  • Cobey SW, Sheppard WS, Tarpy DR (2012) Status of Breeding Practices and Genetic Diversity in Domestic U.S. Honey Bees. In: Honey Bee Colony Health: Challenges and Sustainable Solutions(eds Sammataro D & Yoder JA), pp. 2536. CRC Press, Boca Raton, FL.
  • Cornuet JM, Luikart G (1996) Description and power analysis of two tests for detecting recent population bottlenecks from allele frequency data. Genetics, 144, 20012014.
  • Cresswell JE, Desneux N, vanEngelsdorp D (2012) Dietary traces of neonicotinoid pesticides as a cause of population declines in honey bees: an evaluation by Hill's epidemiological criteria. Pest Management Science, 68, 819827.
  • Currie RW, Pernal SF, Guzman-Novoa E (2010) Honey bee colony losses in Canada. Journal of Apicultural Research, 49, 104106.
  • Dall'Olio R, Marino A, Lodesani M, Moritz RFA (2007) Genetic characterization of Italian honeybees, Apis mellifera ligustica, based on microsatellite DNA polymorphisms. Apidologie, 38, 207217.
  • De la Rua P, Galian J, Serrano J, Moritz RFA (2001) Genetic structure and distinctness of Apis mellifera L. populations from the Canary Islands. Molecular Ecology, 10, 17331742.
  • De la Rua P, Jaffe R, Dall'Olio R, Munoz I, Serrano J (2009) Biodiversity, conservation and current threats to European honeybees. Apidologie, 40, 263284.
  • De la Rúa P, Jaffé R, Muñoz I, Serrano J, Moritz RFA (2013) Conserving genetic diversity in the honey bee. Comments on Harpur et al. (2012). Molecular Ecology.
  • Dlugosch KM, Parker IM (2008) Founding events in species invasions: genetic variation, adaptive evolution, and the role of multiple introductions. Molecular Ecology, 17, 431449.
  • vanEngelsdorp D, Evans JD, Saegerman C et al. (2009) Colony collapse disorder: a descriptive study. PLoS ONE, 4, e6481. doi:10.1371/journal.pone.0006481
  • Franck P, Garnery L, Celebrano G, Solignac M, Cornuet JM (2000) Hybrid origins of honeybees from Italy (Apis mellifera ligustica) and Sicily (A-m. sicula). Molecular Ecology, 9, 907921.
  • Garnery L, Franck P, Baudry E et al. (1998) Genetic diversity of the west European honey bee (Apis mellifera mellifera and A. m. iberica). II. Microsatellite loci. Genetics Selection Evolution, 30, S49S74.
  • Garrigan D, Hammer MF (2006) Reconstructing human origins in the genomic era. Nature Reviews Genetics, 7, 669680.
  • Han F, Wallberg A, Webster MT (2012) From where did the Western honeybee (Apis mellifera) originate? Ecology and Evolution, 2, 19491957.
  • Harpur BA, Zayed A (2013) Accelerated evolution of innate immunity proteins in social insects: adaptive evolution or relaxed constraint? Molecular Biology and Evolution, doi:10.1093/molbev/mst061. [Epub ahead of print].
  • Harpur BA, Minaei S, Kent C, Zayed A (2012) Management increases genetic diversity of honey bees via admixture. Molecular Ecology, 21, 44144421.
  • Harpur BA, Sobhani M, Zayed A (2013) A review of the consequences of complementary sex determination and diploid male production on mating failures in the Hymenoptera. Entomologia Experimentalis et Applicata, 146, 156164.
  • Heimpel GE, de Boer JG (2008) Sex determination in the Hymenoptera. Annual Review of Entomology, 53, 209230.
  • Jensen AB, Palmer KA, Boomsma JJ, Pedersen BV (2005) Varying degrees of Apis mellifera ligustica introgression in protected populations of the black honeybee, Apis mellifera mellifera, in northwest Europe. Molecular Ecology, 14, 93106.
  • Kent CF, Issa A, Bunting AC, Zayed A (2011) Adaptive evolution of a key gene affecting queen and worker traits in the honey bee, Apis mellifera. Molecular Ecology, 20, 52265235.
  • Kent CF, Minaei S, Harpur BA, Zayed A (2012) Recombination is associated with the evolution of genome structure and worker behavior in honey bees. Proceedings of the National Academy of Sciences of the United States of America, 109, 1801218017.
  • Kolbe JJ, Glor RE, Rodriguez Schettino L et al. (2004) Genetic variation increases during biological invasion by a Cuban lizard. Nature, 431, 177181.
  • Koulianos S, Crozier RH (1996) Mitochondrial DNA sequence data provides further evidence that the honeybees of Kangaroo Island, Australia are of hybrid origin. Apidologie, 27, 165174.
  • Lavergne S, Molofsky J (2007) Increased genetic variation and evolutionary potential drive the success of an invasive grass. Proceedings of the National Academy of Sciences of the United States of America, 104, 38833888.
  • Mattila HR, Seeley TD (2007) Genetic diversity in honey bee colonies enhances productivity and fitness. Science, 317, 362364.
  • Muñoz I, De la Rúa P (2012) Temporal analysis of the genetic diversity in a honey bee mating area of an island population (La Palma, Canary Islands, Spain). Journal of Apicultural Science, 56, 4149.
  • Muñoz I, Madrid-Jiménez MJ, De la Rúa P (2012) Temporal genetic analysis of an introgressed island honey bee population (Tenerife, Canary Islands, Spain). Journal of Apicultural Research, 51, 144146.
  • Oldroyd BP (2012) Domestication of honey bees was associated with expansion of genetic diversity. Molecular Ecology, 21, 44094411.
  • Oldroyd BP, Fewell JH (2007) Genetic diversity promotes homeostasis in insect colonies. Trends in Ecology & Evolution, 22, 408413.
  • Oldroyd BP, Cornuet JM, Rowe D, Rinderer TE, Crozier RH (1995) Racial admixture of Apis-Mellifera in Tasmania, Australia - similarities and differences with natural hybrid zones in Europe. Heredity, 74, 315325.
  • Oleksa A, Chybicki I, Tofilski A, Burczyk J (2011) Nuclear and mitochondrial patterns of introgression into native dark bees (Apis mellifera mellifera) in Poland. Journal of Apicultural Research, 50, 116129.
  • Pritchard JK, Stephens M, Donnelly P (2000) Inference of population structure using multilocus genotype data. Genetics, 155, 945959.
  • Rortais A, Gérard A, Alburaki M, Legout H, Garnery L (2011) Review of the DraI COI-COII test for the conservation of the black honeybee (Apis mellifera mellifera). Conservation Genetics Resources, 3, 383391.
  • Ruttner F (1988) Biogeography and Taxonomy of Honeybees. Springer, New York.
  • Sheppard WS (2012) Managed pollinator CAP coordinated agricultural project a national research and extension initiative to reverse pollinator decline honey bee genetic diversity and breeding - towards the reintroduction of European germ plasm. American Bee Journal, 152, 155158.
  • Soland-Reckeweg G, Heckel G, Neumann P, Fluri P, Excoffier L (2009) Gene flow in admixed populations and implications for the conservation of the Western honeybee, Apis mellifera. Journal of Insect Conservation, 13, 317328.
  • Tarpy DR (2003) Genetic diversity within honeybee colonies prevents severe infections and promotes colony growth. Proceedings of the Royal Society of London. Series B: Biological Sciences, 270, 99103.
  • Whitfield CW, Behura SK, Berlocher SH et al. (2006) Thrice out of Africa: ancient and recent expansions of the Honey Bee, Apis mellifera. Science, 314, 642645.
  • Zayed A (2004) Effective population size in Hymenoptera with complementary sex determination. Heredity, 93, 627630.
  • Zayed A (2009) Bee genetics and conservation. Apidologie, 40, 237262.
  • Zayed A, Packer L (2005) Complementary sex determination substantially increases extinction proneness of haplodiploid populations. Proceedings of the National Academy of Sciences of the United States of America, 102, 1074210746.
  • Zayed A, Robinson GE (2012) Understanding the relationship between brain gene expression and social behavior: lessons from the honey bee. Annual Review of Genetics, 46, 591615.
  • Zayed A, Whitfield CW (2008) A genome-wide signature of positive selection in ancient and recent invasive expansions of the honey bee Apis mellifera. Proceedings of the National Academy of Sciences of the United States of America, 105, 34213426.
  • van der Zee R, Pisa L, Andonov S et al. (2012) Managed honey bee colony losses in Canada, China, Europe, Israel and Turkey, for the winters of 2008–9 and 2009–10. Journal of Apicultural Research, 51, 91114.

B.A.H. is a graduate student whose research focuses on honey bee immunity and genetics. S.M. is also a graduate student with an interest in the molecular evolution. C.K.'s research focuses on the evolutionary genomics of social behaviour. A.Z.'s research focuses on understanding the genetic basis of worker phenotypes, and the evolution of sociality.