In this synthesis we apply coevolutionary models to the interactions between socially parasitic ants and their hosts. Obligate social parasite systems are ideal models for coevolution, because the close phylogenetic relationship between these parasites and their hosts results in similar evolutionary potentials, thus making mutual adaptations in a stepwise fashion especially likely to occur. The evolutionary dynamics of host‐parasite interactions are influenced by a number of parameters, for example the parasite's transmission mode and rate, the genetic structure of host and parasite populations, the antagonists’ migration rates, and the degree of mutual specialisation. For the three types of obligate ant social parasites, queen‐tolerant and queen‐intolerant inquilines and slavemakers, several of these parameters, and thus the evolutionary trajectory, are likely to differ. Because of the fundamental differences in lifestyle between these social parasite systems, coevolution should further select for different traits in the parasites and their hosts. Queen‐tolerant inquilines are true parasites that exert a low selection pressure on their host, because of their rarity and the fact that they do not conduct slave raids to replenish their labour force. Due to their high degree of specialisation and the potential for vertical transmission, coevolutionary theory would predict interactions between these workerless parasites and their hosts to become even more benign over time. Queen‐intolerant inquilines that kill the host queen during colony take‐over are best described as parasitoids, and their reproductive success is limited by the existing worker force of the invaded host nest. These parasites should therefore evolve strategies to best exploit this fixed resource. Slavemaking ants, by contrast, act as parasites only during colony foundation, while their frequent slave raids follow a predator‐prey dynamic. They often exploit a number of host species at a given site, and theory predicts that their associations are best described in terms of a highly antagonistic coevolutionary arms race.

We investigated how individual strategies combine with demographic and ecological factors to determine local and migratory movements in the double-crested cormorants (Phalacrocorax auritus). One hundred and forty-five cormorants were captured from 14 nesting colonies across the Great Lakes area and fitted with satellite transmitters. We first tested the hypotheses that sexual segregation, density-dependent effects, and the intensity of management operations influenced home range size during the breeding season. The influence of these factors appeared to be limited in part due to random variability in foraging and dispersal decisions at individual and colony levels. We also designed a statistical framework to investigate the degree and determinants of migratory connectivity. Our analyses revealed a significant migratory connectivity in cormorants, although we also observed a nonnegligible amount of individual variability and flexibility. Our data were most consistent with the existence of a migratory divide across the Great Lakes, with western populations using mainly the Mississippi Flyway and eastern populations the Atlantic Flyway. Previous and current studies suggest that the divide cannot be explained by past divergence in isolation, a way to diminish travel cost, or the Appalachians constituting an ecological barrier per se but is rather the consequence of the distribution of suitable stopover and nonbreeding areas. However, a parallel migration system and no migratory divide could not be entirely ruled out with present data.

Recently, avian brood parasites and their hosts have emerged as model systems for the study of host–parasite coevolution. However, empirical studies of the highly analogous social parasites, which use the workers of another eusocial species to raise their own young, have never explicitly examined the dynamics of these systems from a coevolutionary perspective. Here, we demonstrate interpopulational variation in behavioural interactions between a socially parasitic slave–maker ant and its host that is consistent with the expectations of host–parasite coevolution. Parasite pressure, as inferred by the size, abundance and raiding frequency of Protomognathus americanus colonies, was highest in a New York population of the host Leptothorax longispinosus and lowest in a West Virginia population. As host–parasite coevolutionary theory would predict, we found that the slave–makers and the hosts from New York were more effective at raiding and defending against raiders, respectively, than were conspecifics from the West Virginia population. Some of these variations in efficacy were brought about by apparently simple shifts in behaviour. These results demonstrate that defence mechanisms against social parasites can evolve, and they give the first indications of the existence of a coevolutionary arms race between a social parasite and its host.