Abstract 1 Trichogramma brassicae Bezdenko is inundatively released against the European corn borer in Switzerland. Because parasitoids dispersing from the release fields might pose a threat to native butterflies, the searching efficiency of T. brassicae was investigated in nontarget habitats. 2 In field studies, T. brassicae was released at rates of 120 000 females/ha. Parasitism of sentinel Ephestia kuehniella egg clusters was 1.6–3.6% in meadows and 2.0–4.0% in flower strips. The respective figures were 57.6–66.7% and 19.2–46.9% in maize, significantly higher than the parasitism rates in the nontarget habitats. Experiments carried out in small field cages confirmed these results: Again, significantly higher parasitism rates were found in maize compared to meadows and flower strips, and also compared to hedgerows (in sleeve cages). 3 To elucidate potential factors underlying the low searching efficiency in nontarget habitats, the behaviour of individual T. brassicae females was investigated on four meadow plants comparatively to maize and a filter paper control. Mean (±SE) walking speed on maize was 2.2 ± 0.2 mm/s, similar to three of the plants tested and filter paper but significantly higher than on Trifolium pratense (0.85 mm/s). A higher turning rate was found on T. pratense, Viola wittrockiana and Plantago lanceolata, in contrast to the longer leaved maize and Alopecurus pratensis. The number of wasps leaving the plant within the observation period differed significantly between plant species, and was twice as high for T. pratense (and the filter paper control) compared to the other plant species. 4 In a choice experiment carried out in a climate cabinet with all five host plant species in cages, we obtained the highest parasitism rates on maize and the lowest parasitism on T. pratense, thus confirming the behavioural observations. 5 In conclusion, there is evidence for a decreased searching efficiency on plants in nontarget habitats compared to maize. However, the data explain only part of the differences found between parasitism in maize compared to nontarget habitats. Other factors, such as the structural complexity of a habitat, may also play a role. We conclude that the risk for butterfly populations in the tested nontarget habitat due to mass released T. brassicae is low.
Swiss Federal Research Station for Agroecology and Agriculture, Reckenholzstrasse 191, 8046 Zürich, Switzerland and 2:
ETH (Swiss Federal Institute of Technology), Institute of Plant Sciences, Applied Entomology, Clausiusstrasse 25/NW, CH-8092 Zürich, Switzerland