Leaf traits are known to influence the preference and performance of phytophagous insects. Traits that influence oviposition can differ from traits that favour larval development, but in native hosts the association between traits usually leads to positive preference-performance relationships. However, when herbivores interact with novel hosts, traits that influence oviposition and successful larval development can become decoupled, leading to poor preference-performance relationships. We investigated traits associated with preference and performance of the horse-chestnut leaf miner, Cameraria ohridella Deschka & Dimic (Lepidoptera: Gracillariidae), on its native host, Aesculus hippocastanum L. (Sapindaceae), and on other species of Aesculus that are novel hosts for the leaf miner. Cameraria ohridella laid eggs on all species of Aesculus, but showed preferences for A. hippocastanum and the phylogenetically closely related Japanese horse-chestnut, Aesculus turbinata Blume. Successful mine development, however, was observed only on A. hippocastanum and A. turbinata, which indicates a partial mismatch between female oviposition and larval performance. The two susceptible Aesculus species were very similar in their leaf traits, which might explain why larvae performed equally well on these species. In contrast, species resistant to C. ohridella [Aesculus chinensis Bunge, Aesculus indica (Camb) Hook, and Aesculus flava Aiton] showed considerable variation in leaf traits, which suggests that different species of Aesculus might have different mechanisms of resistance to C. ohridella, supporting the idea of different defensive syndromes in different species of plants.Read less <