The structure of ant societies. Dispersal, recognition, and population organization in Formica exsecta

Ants evolved to live in families composed of a reproductive queen and her offspring, and close relatedness between family members was central to the evolution of their social way of life. Many ants have since diverged from this ancestral structure. The presence of multiple reproductive queens in a colony is one such departure, and it remains puzzling because it lowers relatedness between nestmates. In the genus Formica, many species are socially polymorphic, with single-queen colonies as well as multi-queen colonies that bud new nests from the parent nest, expanding into networks called supercolonies. I studied the mechanisms underlying differences between social forms in Formica exsecta, a socially polymorphic species, using behavioral, genetic, and chemical methods.

I first investigated behavioral variation and dispersal in young queens. Surprisingly, the social form of the home nest did not predict queen dispersal. Instead, I found evidence of personality-linked dispersal: queens more willing to stand their ground against a simulated predator cue were less likely to disperse. I next studied how workers respond to young queens and males attempting to enter their nests, and again found that social form did not influence worker responses to nestmates or non-nestmates, suggesting that worker discrimination is not the mechanism by which multi-queen, but not single-queen, colonies recruit additional queens. Finally, combining chemical, genetic, and behavioral data, I investigated two large, dense populations whose low internal aggression suggested coherent supercolonies. Contrary to expectation, genetic data showed that these populations were composed mainly of independent single-queen colonies with unusually low baseline aggression, showing that low aggression alone is an unreliable sign of supercoloniality, because behavioral and genetic markers of social organization are not always linked.

Together, my results show that we still understand little about the proximate mechanisms generating variation in queen number across social forms. Behavioral variation among individuals and populations cuts across social forms, suggesting that the suite of traits often associated with multi-queen organization is more loosely coupled than previously thought.