Evolution of social structure in the ant genus Myrmecia fabricius (Hymenoptera: Formicidae)

Eusocial insects vary significantly in colony queen number and mating frequency, resulting in a wide range of social structures. Detailed studies of colony genetic structure are essential to elucidate how various factors affect the relatedness and the sociogenetic organization of colonies. The polygyny-vs.-polyandry hypothesis argues that polygyny and polyandry should be negatively associated since both can result in increased intracolonial genetic variability and have costs. However, evidence for this long-debated hypothesis has been lacking at the intraspecific level. Ants of the genus Myrmecia Fabricius display many ancestral biological traits, and thus are considered valuable in investigating the origin and evolution of more derived social behaviors, life histories and morphologies as found in other ants. In this research, a set of highly polymorphic microsatellite loci were developed, and employed to determine fine-scale sociogenetic organizations in two species of this genus, i.e., the bulldog ant M. brevinoda and the jumper ant M. pilosula. The polygyny-vs.-polyandry hypothesis was also examined using both cases. In addition, I evaluated nestmate recognition in M. brevinoda using behavioural assays, and investigated the impacts of colony social structure, genetic and spatial distances on recognition and aggression. M. brevinoda is facultatively polygynous and polyandrous. The numbers of queens per colony varied from 1 to 6, and queens were inferred to mate with 1 to 10 males. Nestmate queens within polygynous colonies were on average related, but the overall relatedness between queens and their mates was indistinguishable from zero. A lack of genetic isolation by distance among nests indicated the prevalence of independent colony foundation. In accordance with the polygyny-vs.-polyandry hypothesis, the number of queens per colony was significantly negatively associated with the estimated number of matings (Spearman rank correlation R = -0.490, P = 0.028). This study thus provides the rare intraspecific evidence for the polygyny-vs.-polyandry hypothesis. Workers of M. brevinoda were always non-aggressive towards nestmates, but acted either aggressively or non-aggressively towards alien conspecifics, suggesting that they are generally able to discriminate between nestmates and non-nestmates. Mantel tests revealed no significant impact of genetic and spatial distances on nestmate discrimination. Moreover, the data appear to lend no support to the hypothesis that colony social structure (queen number) variation significantly affects nestmate recognition and aggression. Thus, the actual mechanism underlying nestmate recognition in this species remains to be resolved. M. pilosula is also facultatively polygynous and polyandrous. The number of queens per colony ranged from 1 to 4, and queens were inferred to mate with 1-9 males. Nestmate queens within polygynous colonies, and queens and their mates, were generally unrelated. This is the first time that the rare co-occurrence of polygyny and high polyandry has been found in the M. pilosula species group. The isolation-by-distance pattern and the occurrence of polygynous polydomy suggest the occurrence of dependent colony foundation in M. pilosula; however, independent colony foundation may co-occur since queens of this species have fully developed wings and can fly. There is no support for the predicted negative association between polygyny and polyandry in ants. Combining the support from the case of M. brevinoda and the rejective evidence from the case of M. pilosula and other intraspecific studies, I suggest that the high costs of multiple matings and the strong effect of multiple matings on intracolonial genetic diversity may be essential to the negative association between polygyny and polyandry, and that any attempt to empirically test this hypothesis should place emphasis upon these two key underlying aspects.