EVOLUTION OF PLANT-RESISTANCE TO MULTIPLE HERBIVORES - QUANTIFYING DIFFUSE COEVOLUTION

Studies of coevolution in plant-herbivore systems have typically focus ed on tight, pairwise interactions between one herbivore and one host plant species. Diffuse coevolution, by contrast, has received much les s empirical attention, presumably because imprecise definitions of dif fuse coevolution have hindered the development of experimental approac hes for distinguishing between pairwise and diffuse coevolution. Here we provide a definition of diffuse coevolution that leads to three cri teria for the operation of pairwise coevolution: susceptibilities (res istances) to different herbivores are genetically uncorrelated, the pr esence/absence of one herbivore does not affect the amount of damage c aused by other herbivores, and the impact of one herbivore on plant fi tness does not depend on the presence/absence of other herbivores. All three criteria must be satisfied for coevolution to be pairwise; if a ny of them fail, coevolution is diffuse. We then describe an experimen tal design and statistical analysis that permit the partitioning of th e total selection imposed on a plant by a set of herbivores into compo nents representing pairwise and diffuse selection, thus allowing deter mination of whether coevolution is pairwise or diffuse.