The challenge of developing a theory on the evolution of insect host and parasitoid interactions is addressed. Stages in the development of theory are recognized as the identification of pattern in nature, the development of hypotheses to account for pattern, and the further study that will validate or refute perceived patterns and processes. Patterns are identified in the context of vegetational ecological succession. Along this gradient there are an increasing number of parasitoid species per host species and an increasing proportion of more generalized (idiobiont) parasitoid species per host species. As the number of parasitoid species per host herbivore species increases, the expectation of higher mortality inflicted upon host populations increases. This expectation is justified on the basis of attempts at biological control that show a correlation between the number of parasitoid species per host and the probability of successful regulation by parasitoids. A predicted pattern is that, in host-parasitoid interactions along vegetational succession, innate responses to semiochemicals by searching parasitoids will be strong in early succession, and learned responses will predominate in late successional communities. Hypotheses that account for the mechanisms in generating pattern are reviewed, based on the literature, and their applications to a theory of biological control are discussed. (C) 1991 Academic Press, Inc.