Performance of leaf beetle larvae on sympatric host and non-host plants Pierluigi Ballabeni & Martine Rahier Laboratoire d’Ecologie Animale et Entomologie, Institut de Zoologie, Universit´ e de Neuchâtel, CH-2007 Neuchâ- tel, Switzerland Key words: larval performance, trade-off, reaction norms, host range evolution, leaf beetle, Oreina, Chrysomelidae Abstract Studies asking the ability of insects to utilize novel host plants often use novel hosts that are allopatric with the insect population under investigation. However, since the outcomes of species interactions are often site-specific, such studies cannot tell us whether a plant would actually be used by a given insect population if the plant grew sympatrically with it. We therefore performed a quantitative genetics experiment to analyse the performance of larvae of the leaf beetle Oreina elongata Suffrian (Coleoptera: Chrysomelidae, Chrysomelinae) on two host and three non-host plants, collected from a site where insects and plants co-occur in the Western Alps. When raised on the non-host Petasites albus (L.), larvae were able to survive equally well as on the two hosts, Adenostyles alliariae (Gouan) and Cirsium spinosissimum (L.), whereas they did not survive on the two other non-hosts, Peucedanum ostruthium (L.) and Rumex alpinus L. On P. albus, growth rate was slightly lower and development time slightly longer than on the two hosts. We found a genotype by environment interaction only for growth rate but not for development time and survival. However, the shape of the reaction norms of growth rates suggests that it is unlikely that selection could favour the inclusion of P. albus into the host range of the study population. Introduction The vast majority of phytophagous insects are spe- cialised in their use of only few host plant species, mainly belonging to the same family. In this context, the physiology of insects and plants is obviously a very important selective factor in shaping host plant use since phytophagous insects must be adapted to di- gest plant nutrients and to deal with plant defensive secondary compounds. However, several examples are known of insect species that use a range of host plants that is actually narrower than the range of plants on which the physiology alone would allow larvae to de- velop and survive (e.g., Wiklund, 1975; Kibota & Courtney, 1991; Futuyma et al., 1994, 1995; more references in Fox & Lalonde, 1993). On one hand, ecological factors, such as natural enemies, compe- tition or meteorological adversities may concur to restrict an insect’s host range (Futuyma & Peterson, 1985; Denno et al., 1990; Jaenike, 1990). On the other hand, ovipositing females may not accept, or may not be attracted to, the chemical or visual stimuli of some of the plants on which their larvae might perform well (Jaenike, 1985; Fox & Lalonde, 1993). When larvae are preadapted to grow and survive on a wider number of plants than those actually used, then a species’ host range might expand if females oviposited by mistake on a potential host not normally utilised (Feeny, 1991; Jaenike & Papaj, 1992; Larsson & Ekbom, 1995). Moreover, if adults showed fidelity to the novel host on which they developed as larvae, then a new host- specific insect race could potentially evolve (Wood et al., 1999). Oviposition mistakes are well docu- mented for lepidopterans and gall makers, the adults of which are very mobile and fly around searching for plant patches for suitable hosts (references in Larsson & Ekbom, 1995). In addition, mistakes are more likely if females that carry high egg loads end up on novel plants that are chemically similar to the normal hosts (Jaenike, 1990). Studies measuring insect survival and growth on novel hosts are often performed using either labora- Published in Entomologia Experimentalis et Applicata 97, issue 2, 175-181, 2000 which should be used for any reference to this work 1