Abstract
Field trials were established at three European sites (Denmark, Eastern France, SouthWest France) of genetically modified maize (Zea mays L.) expressing the CryIAb Bacillus thuringiensis toxin (Bt), the nearisogenic non-Bt cultivar, another conventional maize cultivar and grass. Soil from Denmark was sampled at sowing (May) and harvest (October) over two years (2002, 2003); from E France at harvest 2002, sowing and harvest 2003; and from SW France at sowing and harvest 2003. Samples were analysed for microbial community structure (2003 samples only) by community-level physiological-profiling (CLPP) and phospholipid fatty acid analysis (PLFA), and protozoa and nematodes in all samples. Individual differences within a site resulted from: greater nematode numbers under grass than maize on three occasions; different nematode populations under the conventional maize cultivars once; and two occasions when there was a reduced protozoan population under Bt maize compared to non-Bt maize. Microbial community structure within the sites only varied with grass compared to maize, with one occurrence of CLPP varying between maize cultivars (Bt versus a conventional cultivar). An overall comparison of Bt versus non-Bt maize across all three sites only revealed differences for nematodes, with a smaller population under the Bt maize. Nematode community structure was different at each site and the Bt effect was not confined to specific nematode taxa. The effect of the Bt maize was small and within the normal variation expected in these agricultural systems.