The most reliable and practicable measure in assuring coexistence in respect to pollen‐mediated gene flow from genetically modified (GM) to conventional maize (Zea mays L.) is an isolation distance separating GM and non‐GM fields. Therefore, we tested distances between 24 and 102 m at three sites in northern Germany using a field orientation representing a worst case scenario concerning wind direction. During the 3 yr of field trials the highest levels of gene flow occurred at the site and year with the longest flowering synchrony and the strongest wind blowing constantly from the GM to the non‐GM field. It was shown that the GM content of a neighboring non‐GM maize field is mainly determined by wind speed and direction as well as by non‐GM maize field depth. Based on the maximum outcrossing data obtained it can be concluded that for non‐GM maize fields being 200 m deep or more an isolation distance of 50 m is sufficient to keep the GM content of the total fields grain harvest below the European Union (EU) labeling threshold of 0.9%. However, non‐GM grain maize fields with smaller field depth require larger isolation distances or additional coexistence measures. In most cases discarding 6 m of the GM maize facing non‐GM maize field edge has proven to be such a valuable measure. In silage maize production 50 m isolation distance is adequate even for non‐GM maize field depths down to 50 m. We recommend flexible separation distances in dependence on non‐GM maize field depth to comply with EU coexistence requirements.

At the field scale, the co-existence of different farming production systems requires strategies to prevent gene flow between adjacent crops. Oilseed rape produces pollen dispersed by wind and insects and the risks of pollen mediated gene flow are significant for this crop. Cleistogamy, the trait of non-opening flowers, could be used to reduce pollen flow. Cleistogamous oilseed rape genotypes were obtained by INRA in France and were bred in order to improve the stability of this trait. In this paper, we examine the reliability of the cleistogamous trait for two inbred lines. The flower opening level was measured at different dates during the flowering period in six field experiments (three sites · 2 years). The results showed that some flowers were partially opened with rates varying from 0.5% to 33% principally depending on genotypes, trials (site and year) and recording dates. Given that other studies have shown that cleistogamy could reduce pollen dispersal, we consider that, even when partially unstable, cleistogamy could be beneficially used in combination with other means in a containment strategy.

Recently, the introduction of GM maize in agricultural production in the EU and elsewhere has raised the issue of adventitious presence of GM seeds in conventional seed lots. Adventitious presence may occur in all arable farming, and at any step in the production of seeds or grain, or in processing of harvested product in the food/feed chain. As of today, there are no official thresholds governing the adventitious presence of GM seeds in conventional seed lots in Europe. However, it is assumed that GM admixture in seed lots could have a considerable influence on the level of adventitious presence in the non-GM harvested product. The experiments highlighted in this paper aim at the consequences of adventitious presence of GM maize seeds in conventional seed lots. It is shown for varieties belonging to the same maturity group that the final GM rate (% seeds) in the harvest product is nearly same as the initial seed admixture (% seeds). This corresponds to Hardy-Weinberg expectations. The variation depends mainly on the flowering coincidence, the site and climatic conditions. In cases where the admixed seeds are of different maturity group, the level of cross-pollination in the harvest product is reduced. Furthermore a comparison between the visual GM seed detection and real-time PCR detection was done. It is evident that the result of the real-time PCR detection method has a more variable uncertainty associated with its results than the visual seed testing method. The accuracy of prediction from % GM seed to % GM DNA depends on the reference material used for calibration curves.