One prominent concern about genetically modified crops is the possibility of environmental impacts from the movement of fitness-enhancing traits to wild plant populations. Decisions to deregulate Bt crops in the USA have relied strongly on arguments that these crops will not interbreed with wild relatives in the permitted growing regions. Limited attention therefore has been directed to analyses of the consequences of gene flow. To provide a transparent evaluation process for risks associated with insecticidal transgene escape, we crafted a series of questions designed to guide this aspect of the risk assessment. We then explored the current knowledge base available for answering such risk-related questions for three Bt crops (cotton, rapeseed, and rice). First, we generated a list of wild relatives of these crops. A definitive list of potential transgene recipients is not yet possible for some crops. Sufficient data are not available for some crops to eliminate certain related plant species from consideration of fertile hybrid formation, thus making lists for these crops subject to speculation. Second, we queried the HOSTS database (UK) to obtain a worldwide listing of lepidopteran species that feed on these crops and their wild relatives, and to determine the host range of the larvae. To our knowledge, this list of 502 lepidopteran species is the first such list published for these crops and wild crop relatives. Third, we used a data set maintained by the Canadian Forest Service to assess Bt toxin susceptibility for these lepidopterans. Only 3% of those species have been tested for susceptibility; and the literature suggests that generalizations about susceptibility among taxa are difficult due to the variability within families. Fourth, we consulted the literature to interpret what is known about the ability of lepidopterans to regulate plant fitness or invasiveness. We could not eliminate the possibility of ecological release due to plant resistance against lepidopterans. In fact, there is strong experimental evidence that lepidopteran herbivores do limit the distribution and/or abundances of at least some wild plant species. Neither could we eliminate the possibility that non-target lepidopterans might have important functions in the ecosystem as pollinators or alternate hosts to natural enemies of pest species. This study suggests that crucial data are lacking for the development of a credible scientific basis to confirm or deny environmental risks associated with the escape of Bt transgene constructs to wild relatives. Given the absence of information on the identity, level of susceptibility, and ecological roles of lepidopterans exploiting specific wild relatives of Bt crops, we suggest that new efforts be directed to assessing possible consequences of lepidopteran mortality on resistant wild relatives.

Pest management can have substantial impacts on non-target species both within and outside the units being managed. Assessment of these impacts is hampered by the lack of even the most basic checklist of the species present in most systems. The maize agroecosytem is of particular interest because of the large area covered and the intensity of widely varying forms of pest management. In this study a database of lepidopteran species that occur within the maize agroecosystem in the United States was compiled. The process was initiated by developing a list of plants present in maize using published sources and the first-hand knowledge of “weed” experts. This list of plant species associated with maize was then cross-listed with lepidopteran host feeding records using published sources. Finally, phenological profiles and conservation rankings were added. Although our list is not exhaustive, we found 132 plant species in 33 families associated with maize, and 229 lepidopteran species in 21 families that feed on these plants. The database of plants and lepidopteran species can be a starting point for assessment of risk to non-target Lepidoptera in maize from chemical control, biological control, and the use of transgenic Bt maize. The lepidopteran species associated with maize were found to be significantly less imperiled, as measured by their conservation rankings, than lepidopteran species as a whole in all habitats. This finding suggests that rare or endangered lepidopteran species are unlikely to be impacted by pest management in maize. Based on the likely lack of impact of pest management in maize on individual species, future studies should focus on potential impacts on the ecological services that lepidopteran species provide.

Release of transgenic insect-resistant crops creates the potential not only for the insect pest to evolve resistance but for the escape of transgenes that may confer novel or enhanced fitness-related traits through hybridization with their wild relatives. The differential response of diamondback moth (Plutella xylostella) populations in eastern and western Canada to Bt-producing (GT) Brassica napus and the potential for enhanced fitness of GT B. napus and weedy GT Brassica rapa × B. napus hybrid populations (F1, BC1, BC2) were studied. Comparative bioassays using neonates and 4th instars showed that GT B. napus and GT B. rapa × B. napus hybrids are lethal to larvae from both populations. No measurable plant fitness advantage (reproductive dry weight) was observed for GT B. napus (crop) and GT B. rapa × B. napus hybrid populations at low insect pressure (1 larva per leaf). At high insect densities (>10 larvae per leaf), vegetative plant weight was not significantly different for GT B. napus and non-GT B. napus, whereas reproductive plant weight and proportion of reproductive material were significantly higher in GT B. napus. Establishment of the Bt trait in wild B. rapa populations may also increase its competitive advantage under high insect pressure.

A misprint occurred in this article: The graphs from Figure 1 and Figure 2 should be exchanged with the graphs from Figure 3 and Figure 4, respectively. The corresponding legends are in the correct position and should not be moved.