Did group selection or kin selection build the beehive?

Biologists still argue about it, and the argument shows how hard it is to explain a sterile worker bee. Darwin called the sterile castes of insects "one special difficulty" that at first seemed "insuperable, and actually fatal to my theory". He anticipated the answer lay in close family relationship, which W. D. Hamilton quantified in 1964 as inclusive fitness theory: natural selection favours helping relatives when the cost of helping is less than the benefit multiplied by the fraction of genes shared.

In ants, bees and wasps there was a neat extra twist. Females develop from fertilised eggs and males from unfertilised ones, so sisters share 75% of their genes with each other, more than they would with their own offspring. Hamilton called them "supersisters", and the idea was often invoked to explain at least nine separate evolutions of eusociality in the Hymenoptera.

Then came the doubts. Robert Trivers noted that those same females share only 25% of their genes with their brothers, so average relatedness to a sibling is 50%. And haplodiploidy is neither necessary nor sufficient for eusociality: it is not the whole story.

In 2010 Martin Nowak, Corina Tarnita and E. O. Wilson argued in Nature for multilevel selection, including group selection, to correct what they saw as deficits in the explanatory power of inclusive fitness. In reply, 137 other evolutionary biologists wrote that those arguments rested on a misunderstanding of evolutionary theory and a misrepresentation of the empirical literature. So kin selection kept many defenders, but the question is not closed, and one view holds that the two are not rival processes at all: the effects of group-level selection can already be captured by Hamilton's rule, if relatedness is defined more broadly, as Wilson himself proposed.