INTERACTIONS BETWEEN HERBIVOROUS INSECTS AND PLANT-INSECT COEVOLUTION

Two alternative paradigms exist for coevolutionary processes in plant-herbivore systems: ''pairwise'' and ''diffuse.'' There is little empirical evidence indicating which mode is more common, in part, we believe, because of an absence of explicit criteria by which the two modes can be distinguished. We present such a criterion, which states that, for selection imposed by different herbivores to be independent (pairwise coevolution), not only must there be no genetic correlation among plant resistances to the different species of herbivores, but also there must be no ecological interactions whereby one herbivore affects the evolution of resistance to other species. We present a model that demonstrates the validity of this criterion. We tested whether this criterion is satisfied by the annual morning glory Ipomoea purpurea and two of its insect herbivores, Deloyala guttata and Bedellia somnulentella, by conducting both greenhouse and field experiments. We placed various densities of the two insect species on plants in a factorial design and tested for a statistical interaction between densities of the two species on plant fitness. We did not detect an interaction in either experiment. Our results suggest, therefore, that selection exerted by insect herbivores on resistance in I. purpurea is pairwise rather than diffuse.