RELATIONS AMONG ASSEMBLAGE SIZE, HOST SPECIALIZATION, AND CLIMATIC VARIABILITY IN NORTH-AMERICAN PARASITOID COMMUNITIES

We classified the parasitoids attacking 628 species of North American herbivorous insects as idiobionts (those that permanently disable their hosts during attack) or koinobionts (those that permit continued host development following parasitization). Using these physiological attributes of parasitoids as estimators of host range, we tested three hypotheses that might explain why average parasitoid assemblage size decreases toward the tropics in complexes associated with exophytically feeding herbivores, whereas it does not in complexes associated with endophytic hosts. We found that the geographical patterns for generalist (i.e., idiobiont) and specialist (i.e., koinobiont) hymenopterous parasitoids and Diptera do not support a host-predation hypothesis or a parasitoid-predation hypothesis and are only partly in accord with the resource-fragmentation hypothesis. We conclude that a modified expression of the resource-fragmentation hypothesis can account for some of the geographical patterns of parasitoid assemblage size, but other forces, as yet unidentified, are also operating.