Maize pollen mediated gene flow in the Po valley (Italy): Source-recipient distance and effect of flowering time

Gene flow in maize can be monitored by measuring the cross-fertilization rate from a pollen source to a pollen recipient plot. According to European Commission Recommendation 2003/556, co-existence measures should allow non-GM crops to be grown and marketed so that the adventitious presence of GM material does not exceed the labelling threshold of 0.9% set by EC Regulation 1829/2003. Using dominant phenotypic markers we have investigated in farm scale fields in the Po Valley (Italy) the effect of: distance between the pollen source and recipient plants, with and without pollen competition; wind; synchrony in flowering times, in determining cross-fertilization. To this purpose, three types of experimental fields were designed: in type 1, a block of pollen source was planted in the middle of a recipient field; in fields of type 2, the source was separated from the recipient maize by fallow soil and/or maize buffer zones of variable shape and dimension; in type 3 experiments, the pollen source was planted within a recipient field of maize hybrids having different growing cycle lengths (and, hence, differing flowering synchrony). The following conclusions could be drawn: (1) the 0.9% cross-fertilization threshold was reached within, on average, 10 m in type I experiments (but exceptionally at 25 m); 17.5 m in type 2a experiments; 1.5 in for areas contiguous to pollen source or to recipient in type 2b experiments; (2) the influence of wind was minor compared to distance between pollen source and recipient; (3) buffer maize plants that shed non source pollen, rather than fallow land, were the most efficient barrier against cross-fertilization. Type 3 experiments allowed to conclude that: (1) little or no reduction in pollen flow was observed if there were only up to 3 days of difference in flowering time between pollen source and recipient; (2) when the time interval was 4-5 days a 25% reduction of pollen flow was recorded; (3) when the time interval was 6 days, the reduction was 50%, reaching levels close to 0% when the off-set was higher than 7 days. (c) 2007 Elsevier B.V. All rights reserved.