Assessment of effects of Bt-oilseed rape on large white butterfly (Pieris brassicae) in natural habitats

Much effort has been allocated to the definition of risk, relevant for the assessment of genetically modified plants. However, few studies have emphasised the limitations in testing methods. In this study, tests for and effects on non-target herbivores were exemplified and evaluated for Pieris brassicae (L.) (Lepidoptera: Pieridae) and a genetically modified Brassica napus L. (Brassicaceae) expressing the Bacillus thuringiensis (Bt)-toxin Cry1Ac. It was established that this herbivore recognises and accepts the transgenic plant as a host. It was found that ovipositing females of P. brassicae preferred the transgenic variety for egg-laying. Therefore, effects of the transgenic host plant on the herbivore were determined. Larvae feeding on the Bt-plants experienced 100% mortality for all larval stages. Based on these observations, a population model was established. The model showed that larval survival is increased with amount of food (number of plants) and reduced with the frequency of transgenic specimens, number of host plants needed for completing larval development, and number of egg-laying butterflies. Such models may both aid the design of further tests for effects and support the assessment whether population effects are likely to occur due to the presence of insect-resistant plants outside the agricultural area.