Monday, January 16, 2006

Inclusive fitness, yes!   posted by Razib @ 1/16/2006 11:34:00 PM

I've made a few comments about inclusive fitness/kin selection that have expressed caution recently, but this paper in Molecular Ecology points in the other direction and reaffirms the power of W.D. Hamilton's theoretical framework. But one must remember that some of the review literature suggests that kin selection might have been a sufficient condition for the initial evolution of eusociality enabled by haplodiploidy, it may not be a necessary one for its persistence. The ideal model holds that since males are haploid, the coefficient of relatedness between sisters who share a father and mother will be 3/4, greater than the 1/2 coefficient of relatedness between mothers and their daughters. Following Hamilton's Rule, B > C/r, where B ~ benefit, C ~ cost and r is coefficient of relatedness, you get the perfect recipe for altruistic behavior driven purely be gene level selection. The reality though is that sisters within many hymenoptera species have been found to have been fathered by different males (reducing their relatedness depending on how related the males were), and some eusocial insects may even have multiple queens who are not not related resident in the colony generating unrelated progeny. This diversity across the taxon, to the point where B > C/r might not hold on average across a colony, suggests that other factors are at play. One hypothesis is that once eusociality evolved via kin selection the proximate mechanisms which fostered and enabled it eventually became efficient enough to perpetuate eusociality (the standard idea of tit-for-tat reciprocal altruism usually is considered for more "intelligent" taxa). In other words, the eusociality enabled by kin selection altered the characteristics of the species to forgo the necessity of inclusive fitness as an all encompassing umbrella as new adaptive vistas were brought into view by the initial behavioral innovation. The point is that evolutionary biology is complicated. A weakness within a model does not imply that one should discard the model as totally lacking in value, rather, one should proceed judiciously.