The Evolution of Genetic Odor-Cue Diversity in Social Hymenoptera

Recognition in many taxa is mediated by genetic-cue diversity. The present study investigates the likely effects of two recognition contexts on genetic odor-cue diversity in social Hymenoptera by means of single-locus, two-allele, and multiple-allele population-genetics models. Full-sister versus half-sister discrimination in polyandrous societies, as in queen-rearing discrimination in the honey bee, leads to reduced allelic diversity when queens mate with more than two males. In doubly mated species, discrimination may raise or lower allelic diversity, depending on the rules used by workers in discrimination, but most plausibly it results in a reduction in allelic diversity to two alleles. In contrast, colony recognition based on genetic odors can increase allelic diversity, if colonies with rare odors produce more reproductives. This assumption seems likely for social insects that use genetic cues for nest-mate recognition. This result is the reverse of one obtained earlier in an analysis of genetic recognition cues in marine invertebrates. This discrepancy occurs because opposite fitness assumptions were made as a result of differences in ecology. The idea that existing allelic diversity at the sex-determination locus can be used in a system of signaling to prevent matched matings is also considered. Signaling may evolve when both queens and males benefit from avoiding matched matings, such as in the fire ant Solenopsis invicta, but not in the honey bee, Apis mellifera, or the stingless bee Melipona quadrifasciata. However, the high incidence of matched matings in the fire ant suggests that signaling does not occur, although signaling cannot be ruled out without additional information about the random expectation of matched matings. Signaling, which depends on coordinated adaptations of cue production and perception, may not occur because it is disadvantageous when rare, on account of small costs in the component adaptations. These results suggest that intercolonial recognition is the most plausible means by which recognition itself can generate genetic odor-cue diversity. However, in species such as the honey bee, in which intercolonial recognition appears primarily environmental, the effect may be small. This leads to the possibility that the weak full-sister versus half-sister discrimination shown by honey bees is partly due to a low diversity of odor cues. Low diversity of odor cues may also facilitate the evolution of highly cooperative worker behavior by making queen-rearing discrimination less worthwhile.