Estimating the rate of gamergate turnover in the queenless ant Diacamma cyaneiventre using a maximum likelihood model

Abstract.

In monogynous ants, the death of the queen generally implies the death of the colony. However, queen replacement by either unrelated or related queen has been described in few species. In queenless ants from the genus Diacamma, the single reproductive worker (gamergate) is replaced by one of her daughters (or occasionally by a sister). From a long-term genetic survey of nests of D. cyaneiventre, we estimated the rate of gamergate turnover as well as the lifespan of workers and gamergate tenure using a maximum likelihood model developed for this purpose. We specifically compared the genotypes of two cohorts of workers sampled at 2 and 16 months interval from the same nests, using five microsatellite markers. To improve the accuracy of the estimates, we also used in the model the nests from the same population sampled only once and analysed by André et al. (2001). The model indicates that the possibility of the same nest not sheltering the same colony at two different sampling dates (colony turnover) was not significantly different from zero in our sample. The likelihood of the model was maximal for a probability of gamergate change p_γ =  0.005 per day (i.e. a gamergate tenure of 200 days) and a worker lifespan w=60 days, indicating that the gamergate ‘s tenure is about 3 times longer than workers’ expected lifespan in the population studied. Moreover, the genetic analysis of the gamergate and brood in three colonies excavated completely, reveals that colony fission can occur just after a gamergate replacement with the sister of the new gamergate reproducing in the new propagule.