On suicidal punishment among Acromyrmex versicolor co-foundresses: the disadvantage in personal advantage

Co-founding queens (‘co-foundresses’) of the obligate desert fungus garden ant Acromyrmex versicolor exhibit a forager specialist who subsumes all foraging risk prior to first worker eclosion (Rissing et al., 1989); laboratory observations suggest a ‘coordination’ mechanism (sensu Aumann, 1974; Rosenthal, 1974; cf. Pollock, 1994b, 1996) that assigns task-specific risk without contention among (unrelated) co-foundresses (Rissing et al., 1989, 1996). In an experiment designed to mimic a ‘cheater’ who refuses a foraging assignment when it is her lot, co-foundresses delayed/failed to replace their forager, usually leading to demise of their garden (Rissing et al., 1996). Such behaviour is ‘suicidal’, as the essential fungus garden cannot be replaced. Failing to replace a cheater is here an extreme punishment where the punisher harms herself to harm the cheater (after Radner, 1980). We simulate co-foundress allocation of foraging risk with haploid, asexual genotypes that either replace a cheater (at some point, the personally best response) or not (suicidal punishment), under both foundress viscosity (likely for A. versicolor) and random assortment. Simulation ‘fair contests’, with populations initially a 50 : 50 mix of potential replacers and punishers, exposed to mutant cheaters across evolutionary time, reveal (latent) suicidal punishment to be superior (under viscosity) to replacing a cheater. While replacing is a personally best response to cheating, viscosity correlates the descendants of replacers and cheaters, ultimately eliminating the former. Suicidal punishment, however, removes cheaters before they reproduce; under viscosity, patches of latent punishers can be quite robust, removing phenotypic cheaters as they appear, relative to patches of latent replacers, which rather incubate more cheaters by assuming the risky foraging task for them. In particular, simulations of A. versicolor population ecology reveal (latent) punishment to be superior, supporting laboratory evidence for suicidal punishment in this species (Rissing et al., 1996).