JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.. Ecological Society of America is collaborating with JSTOR to digitize, preserve and extend access to Ecology.Abstract. Comparisons of traits of outbreaking and nonoutbreaking leaf-eating Lepidoptera and Symphyta have shown that spring-feeding species are more likely to have outbreaks than are summer-feeding species. It has been suggested that variable synchrony with host budburst causes the population sizes of spring-feeding species to be more variable primarily because of the negative effects of older leaves on insects. While much evidence exists that leaf age can directly affect survival and reproduction of insects, few studies have looked at the population-level effects of variable phenology, and especially the potential for complex direct and indirect interactions with natural-enemy effects. To examine the consequences of variable phenology for population growth of an outbreaking insect, we manipulated the timing of gypsy moth (Lymantria dispar) egg hatch in the field. We released large numbers of gypsy moth larvae into replicate plots in the field at three times relative to budburst. Survival of larvae in protective sleeves was high unless they were released very long before host budburst, and leaf age had a direct negative effect on fecundity: the later the release, the lower was the fecundity of insects reared on white oak and black oak. The later the release, however, the greater was the dispersal by ballooning, which had the effect of reducing local densities. Because the most important natural enemies impose density-dependent mortality, dispersal had the effect of raising survival rate for later release dates. The direct host-plant effect and natural-enemy effects exhibited opposing influences on population processes, because fecundity decreased with release date but survival increased. The survival advantage of the late release outweighed the loss in fecundity, so that the expected population growth rate was highest for the latest release. The net effects of phenology on insect population growth thus depend largely on natural-enemy effects.Very different conclusions would have been drawn had we measured mortality only in protective sleeves and not in the presence of natural enemies. The strong natural-enemy effects may explain the large variability in outcome of plant-herbivore interactions and contribute to the high variability in population size of spring-feeding species.