Selection of and performance on host plants by specialist herbivores is thought to be governed by host plant chemistry. The close chemical relationship between specialists and their hosts is the basis of the phytochemical co-evolution theory, which postulates that selection pressure by specialists drives the evolution of both the diversity in structural classes of plant secondary metabolites, as well as the diversity in closely related chemical structures within the major classes. However there is currently little evidence to suggest that specialists can detect small structural differences in secondary metabolites of their hosts, or that such differences affect host choice or performance of specialists. We set out to test this hypothesis by conducting no-choice and choice tests using seventeen plant species of the genus Senecio (Asteraceae) and one outgroup (Boraginaceae) containing pyrrolizidine alkaloids (PA), and four PA- sequestering specialist herbivores of the genus Longitarsus (Coleoptera, Chrysomelidae). We also assessed whether variation in feeding by specialist herbivores is partially attributable to different resource use strategies of the plants. We found no evidence that intra-generic differences in PA profiles affect feeding by Longitarsus beetles to a significant extent, neither when using the original nor the phylogenetically corrected data set. In contrast, our results indicate that host-choice was related to plant dry matter content. These findings imply that selection pressure by specialist herbivores is not likely a driving force in evolutionary diversification within the PAs.