Non-native plants may benefit, briefly or permanently, from natural enemy release in their invaded range, or may form novel interactions with native enemy species. Likewise, newly arrived herbivores may develop novel associations with native plants or, where their hosts have arrived ahead of them, re-establish interactions that existed previously in their ancestral ranges. Predicting outcomes from this diversity of novel and re-established interactions between plants and their herbivores presents a major challenge for invasion biology. We report on interactions between the recently arrived invasive planthopper Prokelisia marginata, and the multi-ploidy Spartina complex of four native and introduced species in Britain, each representing a different level of shared evolutionary history with the herbivore. As predicted, S. alterniflora, the ancestral host, was least impacted by planthopper herbivory, with the previously unexposed native S. maritima, a nationally threatened species, suffering the greatest impacts on leaf length gain, new leaf growth and relative water content. Contrary to expectations, glasshouse trials showed P. marginata to preferentially oviposit on the invasive allododecaploid S. anglica, on which it achieved earlier egg hatch, faster nymphal development, larger female body size and greatest final population size. We suggest P. marginata is in the process of rapid adaptation to maximise its performance on what is now the most abundant and widespread host in Britain. The diversity of novel and re-established interactions of the herbivore with this multi-ploidy complex makes this a highly valuable system for the study of the evolutionary ecology of plant–insect interactions and their influence on invasion dynamics.