9420319 Feeny Most plant-feeding insects have specialized preferences, feeding and laying their eggs on only a few plant species. They find their host plants largely by smell or taste of chemical compounds produced by the plant. Different plant species attacked by the same insect often share one or more kinds of unusual chemical compounds, even when the host plants themselves are not closely related to one another. This suggests that the ability of insects to add new plant species to their diet is facilitated by the shared occurrence of compounds that the insects use for cues. The goal of this research is to find out how chemical similarities and differences among plants have influenced present-day associations between different plant species and the insects that feed on them. The swallowtail butterflies are used as a model system in this research since their evolutionary relationships and larval host plants are especially well known. The research will identify the mixtures of plant compounds used as egg-laying stimulants by four insect species that differ greatly in the kinds of plants they attack. The data will be used to test the hypotheses that plant selection by insects is triggered by relatively fixed sensory responses to particular compounds, and that movements to new host species are unlikely unless those species produce the same compounds. The results will provide valuable insights into the degree to which sensory responses by insects to plant chemicals are conservative rather than labile over evolutionary time. Knowledge of both the cues insects currently use to locate host plants and the frequency with which insects abandon those cues to take advantage of new hosts has obvious implications for the wise deployment of new crop varieties whose chemical profiles have been or can be modified by genetic engineering.