Group selection, inclusive fitness, and ants ants (Hymenoptera: Formicidae) • 10-12,000 species; 15-25% of animal biomass worldwide • all are eusocial colonies of fertile queens & sterile workers, 3 classes of offspring: • gynes • workers • males • how is eusociality produced and maintained? ants (Hymenoptera: Formicidae) • 10-12,000 species; 15-25% of animal biomass worldwide • all are eusocial colonies of fertile queens & sterile workers, 3 classes of offspring: • gynes • workers • males • how is eusociality produced and maintained? ants (Hymenoptera: Formicidae) • 10-12,000 species; 15-25% of animal biomass worldwide • all are eusocial colonies of fertile queens & sterile workers, 3 classes of offspring: • gynes • workers • males • how is eusociality produced and maintained? inclusive fitness, kin selection • W.D. Hamilton (1964, 1972), then Trivers & Hare (1976), then Boomsma & Grafen (1990, 1991) • fitness = direct reproduction + effects on others’ reproduction • aiding in reproduction of others is favored if: fitness benefit to relative x relatedness > fitness cost to self inclusive fitness, kin selection • W.D. Hamilton (1964, 1972), then Trivers & Hare (1976), then Boomsma & Grafen (1990, 1991) • fitness = direct reproduction + effects on others’ reproduction • aiding in reproduction of others is favored if: fitness benefit to relative x relatedness > fitness cost to self inclusive fitness, kin selection • W.D. Hamilton (1964, 1972), then Trivers & Hare (1976), then Boomsma & Grafen (1990, 1991) • fitness = direct reproduction + effects on others’ reproduction • aiding in reproduction of others is favored if: fitness benefit to relative x relatedness > fitness cost to self ants: haplodiploidy haplodiploidy • due to haplodiploid reproduction, ant workers may be more closely related to a queen’s offspring than to their own • this supports the evolution of eusociality • however, conditions for rqueen’s offspring > rown offspring are limited, and origins of eusociality aren’t tractable haplodiploidy • due to haplodiploid reproduction, ant workers may be more closely related to a queen’s offspring than to their own • this supports the evolution of eusociality • however, conditions for rqueen’s offspring > rown offspring are limited, and origins of eusociality aren’t tractable haplodiploidy • due to haplodiploid reproduction, ant workers may be more closely related to a queen’s offspring than to their own • this supports the evolution of eusociality • however, conditions for rqueen’s offspring > rown offspring are limited • origins of eusociality aren’t tractable haplodiploidy • monogynous colonies & singly mated queens & 1:1 sex investment ratios relatedness of workers to queen’s offspring = 0.5 • if sex investment ratio = relatedness skew (1:3 males:females) r = 0.625 • 1:3 sex investment ratio is generally found in these conditions haplodiploidy • monogynous colonies & singly mated queens & 1:1 sex investment ratios relatedness of workers to queen’s offspring = 0.5 • if sex investment ratio = relatedness skew (1:3 males:females) r = 0.625 • 1:3 sex investment ratio is generally found in these conditions haplodiploidy • monogynous colonies & singly mated queens & 1:1 sex investment ratios relatedness of workers to queen’s offspring = 0.5 • if sex investment ratio = relatedness skew (1:3 males:females) r = 0.625 • 1:3 sex investment ratio is generally found in these conditions problems • polygynous colonies? • multiple mating of queens? • < relatedness of workers to the queen’s offspring • but if you can’t explain eusociality, at least explain sex ratios... problems • polygynous colonies? • multiple mating of queens? • < relatedness of workers to the queen’s offspring • but if you can’t explain eusociality, at least explain sex ratios... problems • polygynous colonies? • multiple mating of queens? • < relatedness of workers to the queen’s offspring • but if you can’t explain eusociality, at least explain sex ratios... multiple mating • multiple mating overall relatedness drops; but relatedness to male offspring is unchanged • more male-biased sex investment ratio multiple mating • multiple mating overall relatedness drops; but relatedness to male offspring is unchanged • more male-biased sex investment ratio multiple mating • effective mating frequency is generally low in ants • facultatively polyandrous ants: predicted changes occur (Leptothorax) or not (Lasius) • obligate polyandry (rare): predicted changes occur (Attini) or not (Pogonomyrmex) multiple mating • effective mating frequency is generally low in ants • facultatively polyandrous ants: predicted changes occur (Leptothorax) or not (Lasius) • obligate polyandry (rare): predicted changes occur (Attini) or not (Pogonomyrmex) multiple mating • effective mating frequency is generally low in ants • facultatively polyandrous ants: predicted changes occur (Leptothorax) or not (Lasius) • obligate polyandry (rare): predicted changes occur (Attini) or not (Pogonomyrmex) polygyny • polygyny relatedness drops • related queens relatedness asymmetry changes • unrelated queens relatedness asymmetry unchanged • polygyny greater cost of gynes? polygyny • polygyny relatedness drops • related queens relatedness asymmetry changes • unrelated queens relatedness asymmetry unchanged • polygyny greater cost of gynes? polygyny • polygyny relatedness drops • related queens relatedness asymmetry changes • unrelated queens relatedness asymmetry unchanged • polygyny greater cost of gynes? polygyny • gynes cost more more male-biased sex ratio (or investment ratio) • related queens more male-biased sex investment ratio • support mixed; polygyne ants generally more male-biased, but: • competitive benefit from neighboring related nests (Linepithema) • polygyny shifts sex investment ratios without relatedness asymmetry changes (e.g. Pheidole, Formica) polygyny • gynes cost more more male-biased sex ratio (or investment ratio) • related queens more male-biased sex investment ratio • support mixed; polygyne ants generally more male-biased, but: • competitive benefit from neighboring related nests (Linepithema) • polygyny shifts sex investment ratios without relatedness asymmetry changes (e.g. Pheidole, Formica) polygyny • gynes cost more more male-biased sex ratio (or investment ratio) • related queens more male-biased sex investment ratio • support mixed; polygyne ants generally more male-biased, but: • competitive benefit from neighboring related nests (Linepithema) • polygyny shifts sex investment ratios without relatedness asymmetry changes (e.g. Pheidole, Formica) split sex ratios • colonies often specialize in production of one sex • Boomsma & Grafen (1990, 1991) colonies specialize in sex to which workers are more related than average • supporting evidence in some taxa, but: • split sex ratios without any relatedness changes (Solenopsis) • or in the opposite direction (e.g., Pheidole) split sex ratios • colonies often specialize in production of one sex • Boomsma & Grafen (1990, 1991) colonies specialize in sex to which workers are more related than average • supporting evidence in some taxa, but: • split sex ratios without any relatedness changes (Solenopsis) • or in the opposite direction (e.g., Pheidole) split sex ratios • colonies often specialize in production of one sex • Boomsma & Grafen (1990, 1991) colonies specialize in sex to which workers are more related than average • supporting evidence in some taxa, but: • split sex ratios without any relatedness changes (Solenopsis) • or in the opposite direction (e.g., Pheidole) nepotism? • polyandry and polygyny selection for nepotism in workers • nepotism is rare or absent! nepotism? • polyandry and polygyny selection for nepotism in workers • nepotism is rare or absent! males? • males generally ignored in inclusive fitness explanations • male fitness increases with female-biased investment • if sex-ratio arguments are correct strong selection in males against multiple mating males? • males generally ignored in inclusive fitness explanations • male fitness increases with female-biased investment • if sex-ratio arguments are correct strong selection in males against multiple mating males? • males generally ignored in inclusive fitness explanations • male fitness increases with female-biased investment • if sex-ratio arguments are correct strong selection in males against multiple mating phylogenetic inertia? • can this explain tolerance of polygyny? • workers may be stuck with sociality, but they are not stuck with polygyny, or even queens: • queen-killing occurs (Linepithema), but not nepotistically! • reproduction can (rarely) be coopted by workers (Rhytidoponera) phylogenetic inertia? • can this explain tolerance of polygyny? • workers may be stuck with sociality, but they are not stuck with polygyny, or even queens: • queen-killing occurs (Linepithema), but not nepotistically! • reproduction can (rarely) be coopted by workers (Rhytidoponera) so what’s the alternative? • reviews of kin selection in Hymenoptera omit competing hypotheses! • what about group selection? so what’s the alternative? • reviews of kin selection in Hymenoptera omit competing hypotheses! • what about group selection? group selection • formed by analogy to natural selection: “This preservation of favourable variations and the rejection of injurious variations, I call Natural Selection.” Darwin, 1859. • Darwin does not specify the units of selection here group selection • formed by analogy to natural selection: “This preservation of favourable variations and the rejection of injurious variations, I call Natural Selection.” Darwin, 1859. • Darwin does not specify the units of selection here group selection • the between-group component of natural selection • preconditions: heritable variation in fitness between groups • this is probably a general property: • random sampling error • assortative group membership group selection • the between-group component of natural selection • preconditions: heritable variation in fitness between groups • this is probably a general property: • random sampling error • assortative group membership group selection • the between-group component of natural selection • preconditions: heritable variation in fitness between groups • this is probably a general property: • random sampling error • assortative group membership group selection • increased recent prominence; E.O. Wilson has announced the demise of kin selection • need not be in conflict with individual-level selection • increases with partitioning of variance between vs. within groups group selection • increased recent prominence; E.O. Wilson has announced the demise of kin selection • need not be in conflict with individual-level selection • increases with partitioning of variance between vs. within groups group selection • increased recent prominence; E.O. Wilson has announced the demise of kin selection • need not be in conflict with individual-level selection • increases with partitioning of variance between vs. within groups what are groups? • all units are groups at a lower level of analysis • the question is: which is most explanatory? • genes of ultimate importance, but selected through phenotype: • which level of phenotype? • primary unit of selection reproduction? • colonies reproduce what are groups? • all units are groups at a lower level of analysis • the question is: which is most explanatory? • genes of ultimate importance, but selected through phenotype: • which level of phenotype? • primary unit of selection reproduction? • colonies reproduce what are groups? • all units are groups at a lower level of analysis • the question is: which is most explanatory? • genes of ultimate importance, but selected through phenotype: • which level of phenotype? • primary unit of selection reproduction? • colonies reproduce what are groups? • all units are groups at a lower level of analysis • the question is: which is most explanatory? • genes of ultimate importance, but selected through phenotype: • which level of phenotype? • primary unit of selection reproduction? • colonies reproduce group selection vs. kin selection • within-group relatedness increases between-group partitioning of variance • the two are compatible explanations, not strict alternatives • but predictions & explanatory value differ group selection vs. kin selection • within-group relatedness increases between-group partitioning of variance • the two are compatible explanations, not strict alternatives • but predictions & explanatory value differ group selection vs. kin selection • within-group relatedness increases between-group partitioning of variance • the two are compatible explanations, not strict alternatives • but predictions & explanatory value differ is relatedness the right metric? • Solenopsis invicta, Linepithema humile, other major invasives characterized by unicoloniality: • low or absent between-nest aggression, often across large areas • relatedness in nests approaches 0 is relatedness the right metric? • Linepithema humile: • although relatedness is ~0, genetic similarity is high • low relatedness may be a measurement artifact •Solenopsis invicta: • between-nest & between-queen cooperation determined by Gp-9 • single-locus identity, and not relatedness, controls social form is relatedness the right metric? • Linepithema humile: • although relatedness is ~0, genetic similarity is high • low relatedness may be a measurement artifact •Solenopsis invicta: • between-nest & between-queen cooperation determined by Gp-9 • single-locus identity, and not relatedness, controls social form do colony characteristics determine success? • Linepithema & Solenopsis: • increased colony size large competitive advantage • Formica: • increased colony size exploitation of concentrated, long-term resources • Pogonomyrmex: decreased relatedness increased colony growth do colony characteristics determine success? • Linepithema & Solenopsis: • increased colony size large competitive advantage • Formica: • increased colony size exploitation of concentrated, long-term resources • Pogonomyrmex: decreased relatedness increased colony growth do colony characteristics determine success? • Linepithema & Solenopsis: • increased colony size large competitive advantage • Formica: • increased colony size exploitation of concentrated, long-term resources • Pogonomyrmex: decreased relatedness increased colony growth explaining inclusive fitness gaps? • absence of nepotism, tolerance of polygyny & low relatedness more efficient colony function, better colony-level performance • split sex ratios may be determined by colony-level factors: • resource limitation • local mate competition • habitat saturation explaining inclusive fitness gaps? • absence of nepotism, tolerance of polygyny & low relatedness more efficient colony function, better colony-level performance • split sex ratios may be determined by colony-level factors: • resource limitation • local mate competition • habitat saturation but: the forgotten variables • cost & benefit terms of Hamilton’s rule generally ignored • can we accomodate everything by hiding it in those variables? • yes, you can! • but there is no explanatory or predictive value • and it obscures explanatory integration but: the forgotten variables • cost & benefit terms of Hamilton’s rule generally ignored • can we accomodate everything by hiding it in those variables? • yes, you can! • but there is no explanatory or predictive value • and it obscures explanatory integration but: the forgotten variables • cost & benefit terms of Hamilton’s rule generally ignored • can we accomodate everything by hiding it in those variables? • yes, you can! • but there is no explanatory or predictive value • and it obscures explanatory integration integration • Pogonomyrmex genetic variation & productivity: • is this analogous to heterozygote advantages? • split sex ratios: • aren’t we looking at evolution of gonochorism, one level up? • do we want to reinvent the wheel for group-level explanations? integration • Pogonomyrmex genetic variation & productivity: • is this analogous to heterozygote advantages? • split sex ratios: • aren’t we looking at evolution of gonochorism, one level up? • do we want to reinvent the wheel for group-level explanations? integration • Pogonomyrmex genetic variation & productivity: • is this analogous to heterozygote advantages? • split sex ratios: • aren’t we looking at evolution of gonochorism, one level up? • do we want to reinvent the wheel for group-level explanations? conclusions • origins of eusociality not tractable in ants • kin selection predictions for sex ratio hold for monogyny + monandry, fail in other cases • kin selection explanations neglect male fitness; predict nepotism; rely on constraints conclusions • origins of eusociality not tractable in ants • kin selection predictions for sex ratio hold for monogyny + monandry, fail in other cases • kin selection explanations neglect male fitness; predict nepotism; rely on constraints conclusions • origins of eusociality not tractable in ants • kin selection predictions for sex ratio hold for monogyny + monandry, fail in other cases • kin selection explanations neglect male fitness; predict nepotism; rely on constraints conclusions • group selection (or multilevel selection) includes kin selection explanations • accounts for variance independent of relatedness • aids explanation of colony functionality, fosters integration of theory across levels conclusions • group selection (or multilevel selection) includes kin selection explanations • accounts for variance independent of relatedness • aids explanation of colony functionality, fosters integration of theory across levels conclusions • group selection (or multilevel selection) includes kin selection explanations • accounts for variance independent of relatedness • aids explanation of colony functionality, fosters integration of theory across levels