Transgenic wheat (Triticum aestivum L.) varieties are being developed and field-tested in various countries. Concerns regarding gene flow from genetically modified (GM) crops to non-GM crops have stimulated research to estimate outcrossing in wheat prior to the release and commercialization of any transgenic cultivars. The aim is to ensure that coexistence of all types of wheat with GM wheat is feasible in accordance with current regulations. The present study describes the result of a field experiment under the semi-arid climate conditions of Madrid, Spain, at two locations ("La Canaleja" and "El Encin" experimental stations) in Madrid over a 3-year period, from 2005 to 2007. The experimental design consisted of a 50 × 50 m wheat pollen source sown with wheat cultivars resistant to the herbicide chlortoluron ('Deganit' and 'Castan' respectively) and three susceptible receptor cultivars ('Abental', 'Altria' and 'Recital') sown in replicated 1 × 1 m plots at different distances (0, 1, 3, 5, 10, 20, 40, 80 and 100 m) and four directions. Outcrossing rates were measured as a percentage of herbicide-resistant hybrids using an herbicide-screening assay. Outcrossing was greatest near the pollen source, averaging 0.029% at 0 m distance at "La Canaleja" and 0.337% at "El Encin", both below the 0.9% European Union regulated threshold, although a maximum outcrossing rate of 3.5% was detected in one recipient plot. These percentages declined rapidly as the distance increased, but hybrids were detected at different rates at distances of up to 100 m, the maximum distance of the experiment. Environmental conditions, as drought in 2004-2005 and 2005-2006, may have influenced the extent of outcrossing. These assays carried out in wheat under semi-arid conditions in Europe provide a more complete assessment of pollen-mediated gene flow in this crop.