Cyclical oligogyny appears to be the best hypothesis for the maintenance of high levels of relatedness among colony members in epiponine wasps, a tribe of Neotropical social wasps characterized by many queens per colony and reproduction by swarms that contain many queens and workers. The cyclical oligogyny hypothesis predicts that queen number declines as colonies age, so colonies about to fission have very few queens, which results in high relatedness among future queens. We investigate key features of this hypothesis by analyzing demographic characteristics of colonies and relatedness in a Venezuelan population of Polybia emaciata. These colonies averaged 30 queens, 276 workers, and 26 males. Relatedness among workers from the same colony of P. emaciata was 0.24 (SE = 0.076). This value can be reconciled with the high numbers of queens only if queens are highly related, as expected under cyclical oligogyny. We found substantial demographic evidence supporting the cyclical oligogyny hypothesis from an examination of numbers of queens, their ages, and eggs per queen. By contrast, we found no support for the hypotheses that relatedness is elevated in epiponine wasps because of inbreeding (f = 0.07) or because of reproductive dominance among queens.