Chance and selection in the evolution of positive species interactions: An ant-plant mutualism and a bacterial commensalism

Mutualisms and commensalisms are ubiquitous and of fundamental ecological importance. However, we lack basic information about the ecological and evolutionary forces involved in the maintenance or transformation of many of these relationships. This dissertation takes two approaches to investigate the ecological genetics of positive species interactions. First, evolutionary change after 1000 generations of coexistence in a bacterial commensalism was documented. Additionally, the ecological and evolutionary ramifications of variation in an ant-plant mutualism were evaluated.; I observed the results of 1000 generations of evolutionary change in a commensalism between two bacteria, Escherichia coli and Enterococcus faecalis. The interaction changed unpredictably into parasitism or commensalism. There was not predictable change in the magnitude or direction of the interaction effects. Control strains that had evolved in isolation also experienced change that affected the interaction, but this background effect did not explain all changes in cohabiting lines. Local coadaptation had little apparent effect, and mutation and drift are likely explanations for differences between lines.; To document the role of natural selection in the interaction between the plant Chamaecrista fasciculata, its herbivores, and its ant mutualists, I measured the relationship between fitness and a mutualistic trait (extrafloral nectar production). The patterns of selection in three environments (unmanipulated control plants, plants treated with insecticide, and plants from which ants had been excluded) were compared. When correlations with plant size were accounted for, I found evidence of stabilizing selection on extrafloral nectar volume consistent with a balance of costs and benefits of nectar production.; Finally, to evaluate the role of variation in nectar production on ant recruitment, I constructed a pump to simulate nectar production at biologically reasonable rates. I used this pump to test the impact of variation in nectar characters on ant recruitment in a natural community of ants. I demonstrated that variation in rate of nectar production, nectar sugar concentration, and nectary number all had an effect on ant recruitment, but that this effect was mediated by preference differences between ant species, and by the spatial position of the plants. Implications of these findings for evolutionary change in the mutualism are discussed.