Reproductive Harmony via Mutual Policing by Workers in Eusocial Hymenoptera

When queens of eusocial Hymenoptera mate two or more times (assuming equal sperm contributions from males and random sperm use), the workers are more closely related to the queen's sons than to the sons of a randomly chosen worker. This suggests that workers should try to prevent other workers from reproducing, and hence producing sons, in species with queens that mate two or more times. It also provides a possible reason for the absence of reproductive workers in many hymenopteran societies. Reproductive harmony may, therefore, and counterintuitively, result from lowered relatedness among workers. Evidence from the biological literature indicates that eusocial Hymenoptera have the necessary behaviors and discriminatory ability to favor queen-produced over worker-produced males, and any such behavior is referred to as "worker policing." Population-genetics simulations of the fate of a "police allele," which confers any marginal increase in policing behavior to the workers carrying it, indicate that such an allele will invade and spread to fixation provided that queens mate two or more times. This conclusion is supported empirically by what is known about queen mating frequency and worker reproduction in the highly eusocial bees and wasps. Societies typically headed by monandrous queens, such as those of the bumblebee (Bombus terrestris) and stingless bees (Meliponinae), have reproductive workers in queenright colonies; those with polyandrous queens, as in honey bees (Apis) and some yellowjacket wasps (Vespula), do not. Bumblebees and yellowjackets are suggested as critical groups for further investigating worker policing, since species with singly and multiply mated queens occur in both. Examination of the cause-effect relationship between queen mating frequency and worker policing indicates that worker policing is caused by queen polyandry but that worker policing is unlikely to cause polyandry, although it may help stabilize it if police workers show behavioral dominance. Thus, mating many times in order to increase worker policing is not envisaged as a potential strategy by which queens win reproductive conflicts with workers. Nepotistically modulated policing (i.e., reduced policing within patrilines) and increased colony-level efficiency (when worker policing increases the colony's integration and harmony) both relax the conditions for the invasion of a police allele. If worker policing increases colony-level efficiency by about 20%, a police allele can always invade even when queens mate only once. Under less stringent but still reasonable conditions, a police allele will invade if worker policing increases colony-level efficiency by only 4.4%. Effective worker policing in colonies in turn relaxes the conditions for selection for self-policing (i.e., selection at the level of the individual workers for forgoing personal reproduction).