Abstract
The potential role of haplodiploid sex determination in promoting the evolution of altruism and eusociality has been the subject of intense debate for over 50 y. Different theoretical models have suggested that haplodiploidy influences relatedness in a way that either does or does not make it easier for altruism to evolve. This debate over the "haplodiploidy hypothesis" can only be resolved with a decisive empirical test that controls for potential phylogenetic bias. Here we critically examine the current state of evidence for an adaptive link between haplodiploidy and eusociality, applying phylogenetically informed methods to ensure that statistical tests reflect independent evolutionary transitions. Using data from 5,678 species, across all major insect orders, we find no evidence that haplodiploidy favors an increased rate of eusocial evolution. We show that this result is robust to: a) different analytical approaches; b) alternative ways of defining both eusociality and haplodiploidy; and c) uncertainty in eusociality assignments. Our analyses suggest that previously reported associations between haplodiploidy and eusociality are likely to have been artifacts, false-positive results primarily driven by a high transition rate to eusociality within the Hymenoptera. This high transition rate could be explained by any factor associated with that group, such as parental care, monogamy, or the possession of a powerful sting.