Background/Question/Methods Insects, despite forming a very large proportion of the terrestrial biota, are generally omitted from conservation planning efforts because patterns of spatial turnover among insect faunas are largely unknown. Because of the high effort associated with insect surveys, there is a need to be able to predict the turnover of insect faunas from the dissimilarity of other, more easily measured, variables. We here report preliminary results investigating a relationship between the dissimilarity of herbivorous insect faunas and the phylogenetic relatedness of their host plants, and the efficacy of this relationship for predicting turnover. Using pyrethrum knockdown techniques, we collected beetles (Coleoptera) from 40 host plant species from four plant families (Ericaceae, Myrtaceae, Proteaceae and Rutaceae). For each plant species, we sampled ten replicates for a total of nine sampling seasons. All sampled plants were located in sclerophyll woodland vegetation in the Sydney region of New South Wales, Australia. Collected beetles were sorted to morphospecies. Phylogenetic relationships of the host plants were determined by constructing a consensus tree drawn from multiple published source trees. Branch lengths were determined using a combination of published node ages (modeled from molecular and fossil evidence) and interpolation. Results/Conclusions Initial data analysis has been conducted on the beetle faunas of two plant families (Ericaceae and Proteaceae), comparing pairwise faunal dissimilarity with patristic distance. Mantel correlation indicates that there is a weak but significant relationship between the turnover of beetle faunas and the phylogenetic relatedness of host plants (r = 0.208, p = 0.003), allowing for the development of a predictive model of this relationship using matrix regression. However, preliminary modeling of rank-transformed pairwise distances indicates that the amount of explained variance is low (r² = 0.043). Ongoing research includes (1) testing for generality of the relationship with other Orders of insects (Hemiptera and Lepidoptera), and other, subtropical and tropical, localities; (2) refining the predictive model by incorporating additional explanatory variables (spatial separation, vegetative characteristics) and iterative function-fitting techniques (as in General Additive Models); and (3) testing the predictions of the model with additional sampling of novel host plant species within previously sampled families.