Conflict resolution and direct benefits in kin-selected conflicts in social groups

Kin selection theory is the leading theory for explaining the evolution of social behaviour in organisms. The theory is based on the application of natural selection theory to the genes that influence social behaviour, where social behaviour includes both cooperation and conflict. It proposes that individuals value one another in proportion to their genetic relatedness, since relatedness measures the chance they share the genes influencing the social behaviour. Because of its fundamental basis, and the need to explain the complexity of animal sociality, investigating kin selection is of major interest to evolutionary ecologists. The theory has proved very successful in explaining social phenomena in some respects, but not in others. Two general reasons for the failures of the theory have been proposed. The first is that social interactions are resolved by as-yet poorly-investigated aspects of the theory, namely, in social conflicts, the relative power held by different sets of individuals, and/or the overall costs to the social group of conflict behaviour. The second explanation is that individuals' social behaviour is driven by direct benefits (gains from personal reproduction), not kin-selected benefits. We propose to test both these ideas using the bumble bee Bombus terrestris as our study organism. In the social Hymenoptera (ants, bees and wasps), kin selection predicts that either queens or workers (which, if reproductive, produce only sons) should monopolise the production of adult male offspring, depending on the relatedness structure of the colony. But the actual level of worker reproduction is often far lower than that predicted. This is the case in B. terrestris. In this species, the theory predicts that workers should produce most of the adult males, but our data show that in fact they produce only c. 5%. We propose to test two hypotheses for the lack of fit between male parentage data and predictions of kin selection theory in B. terrestris. The hypotheses stem, respectively, from the ideas of conflict resolution (through the unequal distribution of power) and direct benefits. They are: (1) most worker-laid eggs are selectively destroyed by the queen or by other workers, with which power over male parentage largely lies; and (2) worker reproduction is not determined by the relatedness of workers to nestmates, as kin selection theory proposes, but is instead a by-product of workers' capacity for intraspecific social parasitism. These hypotheses arise from two sets of observations. The first is that worker-laid eggs in B. terrestris are frequently eaten by the queen or other workers, although the extent of this has never been quantified. The second is our recent discovery that B. terrestris workers enter other colonies of their own species in which they then produce sons as reproductive 'drifters' (intraspecific social parasites). We will test Hypothesis 1 by using observations, experimentation and genetic parentage analyses (based on microsatellite markers) to determine whether the major consumers of worker-laid eggs in B. terrestris are the queen, reproductive workers, or workers as a whole. We will test Hypothesis 2 by using the same techniques to determine whether colonies with the most reproduction by resident workers are those exporting the most reproductive drifter workers, and/or whether reproductive drifters are losers in dominance battles with resident workers. We will also measure the frequency of reproductive drifter workers in field colonies and effects of nest aggregation on drifting behaviour. Our planned work is novel because it addresses new phenomena and unanswered research questions. Its impact will be broad because of the fundamental nature of the research, widespread interest in social evolution, and the possibility that the results will cause a reassessment of long-standing concepts. It should advance considerably our understanding of the evolutionary ecology of social behaviour.