Plant fitness assessment for wild relatives of insect resistant crops.

Risk assessments of new insect-resistant crops will need to estimate the potential for increased weediness of wild crop relatives as a consequence of gene flow. When field experiments are precluded by containment concerns, simulation experiments can identify hazards or measure expected differences between GMOs and parental plants. To measure plant fitness consequences of wild plant protection from Bt-susceptible herbivores, we used topical sprays of bacterial Bacillus thuringiensis larvacide (Bt) on Brassica rapa. Spontaneous crosses between B. rapa and Bt cole crops cannot be precluded, especially if adoption of Bt varieties leads to high exposure. We compared survivorship and seed output of B. rapa that were either protected from or exposed to Bt-susceptible Lepidoptera in the various conditions where hybrids are likely to occur: cultivated (disked) soil, uncultivated agricultural field margins, and nearby non-crop habitats (meadows and ruderal areas). The relative effect of herbivore protection varied among years, habitats, and populations of seedlings. In 2003-2004, Bt sprays did not result in lower herbivory on B. rapa, and plant fitness was not increased. However, in 2004-2006 B. rapa seedlings protected from Bt-susceptible herbivores lived 25% longer, on average, than those that were exposed to these herbivores. In addition, an average B. rapa seedling sprayed with Bt throughout its lifetime was twice as likely to produce siliques and had 251% of the seed output of a seedling exposed to herbivores. The fitness advantage of Bt-based plant protection was apparent in 2004-2005 in half the plants that experienced higher herbivory, and for 2005-2006, was more pronounced in agricultural habitats than in meadows with established, perennial vegetation and less disturbance. Positive effects of Bt-based plant protection and greater fitness in disturbed habitats suggest that crop-wild gene flow may benefit weed populations, and that field tests with herbivore exclusion/addition experiments are feasible alternatives when molecular containment of transgenes restricts field experiments with insect resistant crop-wild hybrids.