Butterflies and plants: preference/performance studies in relation to plant size and the use of intact plants vs. cuttings

Plants have evolved a number of defences to ameliorate herbivore attacks including chemicals induced by mechanical wounding. Such changes in plant chemical composition are potential confounding factors in experiments on plant – insect interactions, which often present cuttings of potential host plants to phytophagous insects. In particular, this could affect studies of female egg‐laying preference and larval performance, because the same plant chemicals that deter certain generalist insects can elevate attacks from more specialized insects. Furthermore, plant cuttings are by definition smaller than intact plants, and any female host size preference could thus affect experiments using plant cuttings. We first assessed female preference and larval performance of a specialist herbivore, Pieris napi (L.) (Lepidoptera: Pieridae, Pierini), confronted with either intact plants or leaf‐cuttings of four Brassicaceae host plants, Alliaria petiolata (Bieb.) Cavara & Grande, Barbarea vulgaris (L.) WT Aiton, Berteroa incana (L.) DC., and Brassica napus (L.). Egg and larval survival did not differ between intact plants and leaf‐cuttings, whereas larval growth was slightly, but significantly, faster on leaf‐cuttings. Females, however, significantly preferred to lay eggs on intact plants of all four hosts, although the preference hierarchy for the intact plants was largely mirrored by that for leaf‐cuttings. We then tested the female preference for different size‐classes of intact B. napus plants. Small individuals received more eggs than larger individuals, and follow‐up experiments showed that this difference was largely generated by a strong female preference for cotyledon leaves; there was no significant difference in female preference for large and small individuals when both carried cotyledons, and females landing on cotyledons were more likely to oviposit compared to when landing on a true leaf. Our study concludes that plant cuttings can serve as adequate proxies for live plants for preference/performance studies, but that experimentalists should be aware of the variation imposed both by plant handling and plant phenology for female oviposition preference.