Kin selection as the key to altruism: its rise and fall

Social Research, Spring, 2005 by Edward O. Wilson

It turns out, however, that this is wrong. Hamilton made three mistakes, which have led to the vitiation of his main thesis concerning altruism and the origin of sociality.

The first mistake, a simple error in arithmetic, is the conclusion that the production of each sister by individuals in haplodiploid societies passes more of their genes to the next generation than does the production of each daughter. As pointed out by Robert L. Trivers in 1976 (Bourke and Franks, 1995), workers also have to raise brothers, and because the gender of these females are determined by haplodiploidy, they are related by only one-fourth of their genes by common descent. In the case of the typical 1:1 gender ratio for the colony as a whole, one-half of males with one-fourth genes in common with their sisters plus one-half of females with three-fourths genes in common with their sisters equals one-half genes in common, so that being altruistic and deferring to the mother queen for reproduction yields the same as having sons and daughters.

In short, the need of a colony to produce males as well as females, or at least the need to adjust the production so as to benefit from a population-of-colonies sex ratio of 1:1, cancels out the gain from raising sisters as opposed to daughters. Yet, as Trivers also pointed out, altruism can still confer an advantage if the workers (all ant workers are female and all males are nonparticipating drones) have control of the colony and invest more in sisters than in brothers in the rearing of the next generation of reproductive females (virgin queens) and males. This choice puts them in conflict with the mother queen, who is equally related to her sons and daughters; hence, in order to maximize the multiplication of her own genes, the queen should opt for a 1:1 sex ratio.

It has come to pass, from abundant field and laboratory studies, that Trivers' predictions have been proved correct. In circumstances where other evidence indicates that workers are in control, the production of reproductives is tilted toward females. When the mother queen is in charge, the ratio of investment centers on 1:1. And when the queen is absent and a worker takes over, the ratio tilts toward males--another inference from the corrected arithmetic.

What Trivers had stumbled across was evidence that queens and their worker daughters are in a situation that promotes internal conflict. In other words, kin selection is dissolutive, at least in part, as opposed to binding, in the social evolution of insects.

The haplodiploid hypothesis, and the seemingly strong evidence that supported it in the 1970s and 1980s, still favored the binding effect of kin selection. But that too has now collapsed. So many phylads have been discovered that contain colonies with altruistic workers--among ambrosia beetles (Kent and Simpson, 1992), for example, snapping shrimps (Duffy, 1996), and even in one species of rodents (Sherman, Jarvis, and Alexander, 1991)--that the association between haplodiploidy and the presence of worker castes is no longer statistically viable. Further, in the gall-making thrips, which like hymenopterans are haplodiploid, males and not just females serve as nonreproductive castes, contrary to the haplodiploid hypothesis of the origins of sterile castes in colonies (Crespi, 1992).