Abstract
The narrow host range of insect herbivores is noted, and some of the reasons why specificity has evolved are reviewed. Works indicating the need for new approaches are pointed out including the possibility that generalist predators provide a suitable pressure. Experiments to test the hypothesis that generalists are more vulnerable than specialists to predators are described. They involve a vespid wasp and over thirty species pairs of caterpillars, matched for size and density. Overall, generalists were taken more readily than specialists: some but not all reasons could be detected. The results are discussed in ecological and evolutionary terms.