Most textbook treatments imply, and almost all theoretical analyses assume, that mutualistic interactions take place between a single pair of interacting partner species. A major goal of this symposium is to broaden and shift this pairwise perspective and make it concordant with the emerging view that locally exclusive mutualisms between just two species are the exception and that many communities include guilds of mutualistic species on one or both sides of the interaction. Many pollination and seed-dispersal mutualisms have long been recognized as diffuse, but recent molecular analyses are revealing unrecognized partner diversity in mutualistic interactions previously thought to be locally species specific. Co-occurring species within a mutualist guild are unlikely to be ecologically equivalent in the way they locate, compete for, and/or reward partners, and so intraguild interactions have the potential to influence population dynamics and patterns of selection in species on both sides of the mutualistic interaction. To illustrate some of these potential complexities for population dynamics, I use simple path analytic models to show that positive pairwise interactions between mutualists do not necessarily translate into positive net interactions within a mutualism involving more than two species. For example, when there is intraguild competition for partners, or even for resources external to the mutualism, the presence of a lower-quality mutualist can negatively affect the partner population by reducing associations it can form with better mutualists. Variation in quality among potential partners is likely to place a premium on traits or behaviors that enhance association with better mutualists. More investigations are needed to determine how variation among interacting mutualists, with respect to characteristics such as longevity, dispersal capability, and competitive ability, influence population dynamics and selection in multispecies mutualisms.