ntroduction of transgenic crops to agriculture has raised concerns about their effects on agro-ecosystems. We compared nine conventional lines of spring wheat with six genetically modified (GM) lines that contained transgenes of resistance against powdery mildew (Pm3b gene) or against fungi in general (Chi and Glu genes). We assessed the persistence and performance of these lines without competition and in experimental weed communities in the field, their seed germination in the laboratory, their survival in fallow plots and effects on post-harvest vegetation in the field.
In competition with weeds, the GM lines showed reduced seed number, plant height and biomass allocation to seeds than their corresponding non-GM lines. No such performance differences were observed without competition. The seedlings of GM lines did not persist longer than those of the corresponding non-GM lines in fallow plots. In the field, GM and non-GM wheat lines had similar performance and persistence and both were able to reproduce in dense weed communities and to survive during winter on fallow plots. Stored in soil in the laboratory, the seeds of GM and non-GM lines either germinated quickly or lost their viability after 3 months. GM and non-GM lines had no differential effects on the structure and diversity of fallow plant communities within the 6-month period of monitoring. Poor seed longevity yet successful plant persistence in weed communities or on fallow plots indicate that not removing a population of growing plants presents a greater risk than allowing the build-up of a soil seed bank regarding the potential escape of transgenic wheat to the environment. Strong varietal differences in persistence point out the importance of case-by-case assessment of new GM varieties and indicate that transgenic traits should preferably be introduced into varieties with low persistence.