A case study of GM maize gene flow in South Africa

Field trials were planted with a central plot of yellow GM maize (0.0576 ha) surrounded by white non-GM maize (13.76 ha), in two different geographic regions over two seasons with temporal and spatial isolations to surrounding commercial maize planting. Cross-pollination from GM to non-GM maize was determined phenotypically across 16 directional transects. Pollen counts during flowering were compared to weather data as well as percentage cross-pollination. The data were transformed logarithmically, and mean percentage cross-pollination was compared to high cross-pollination.Although there was a general congruency between wind data, pollen load and cross-pollination, it is evident that wind data and pollen load do not solely explain the directional extent of cross-pollination and that swirling winds may have contributed to this incongruence. Based on the logarithmic equations of cross-pollination over distance, 45 m is sufficient to minimize cross-pollination to between <1.0% and 0.1%, 145 m for <0.1% to 0.01% and 473 m for <0.01% to 0.001%. However, compared to this, a theoretical isolation distance of 135 m is required to ensure a minimum level of cross-pollination between <1.0% and 0.1%, 503 m for <0.1% to 0.01% and 1.8 km for <0.01% to 0.001% based on high values of cross-pollination.Based on the results of this study, the use of mean values of cross-pollination over distance may result in an underestimation of gene flow. Where stringent control of gene flow is required, for example, for non-GM seed production or for GM field trials under contained use, the high values of cross-pollination should be used to determine isolation distance. However, this may not be practical in terms of the isolation distance required. We therefore suggest that temporal and distance isolations be combined, taking into account the GM maize pollen sources within the radius of the most stringent isolation distance required.South Africa is one of the few African countries that have intro