The biological and environmental factors that determine cross-fertilisation to fields of oilseed rape varieties of a range of male fertility will be measured. Particular attention will be given to statistical rigour in the estimates of whole-field cross-fertilisation. The proposal arises out of concerns that crossing from surrounding fields might be much higher to fields of many modern varieties of oilseed rape, in which only a proportion (variously 20 to 90%) of the plants produce pollen, than to fully fertile varieties in which all plants produce their own pollen. Research on fully fertile types has shown that the whole-field value of cross-fertilisation from one field to another nearby is generally in the range 0.1-0.3%. There is little information on the comparative value for partly fertile varieties (e.g. partially restored hybrids and varietal associations), but early trials and theoretical estimates indicate a much higher crossing percentage to these varieties. Given the need to ensure high purity of a crop type at harvest, research is now required to estimate whole-field cross fertilisation in these varieties. Attention should be given particularly to the mechanisms and pollen-transfer vectors responsible (notably the relative roles of wind and insects) and to the contribution of regional pollen sources, which are unlikely to be negligible for these varieties. The methodologies for the statistical approach, for detecting low frequency crossing and for measuring point to point transfer of pollen both from nearby and more widely in the surrounding area will be established in the first year; subsequently statistical, biologically based, models will be developed to guide experiments and ultimately to predict cross-fertilisation. A consistent approach will then be applied in the second year to estimating, at two study sites, whole-field crossing from a source field to receptor fields close to the source and around 1 km away. The purpose of this configuration is to separate local pollen transfer, which is likely to be manifest as a steep decline in crossing from the facing edge of a receptor field near the source, from regional pollen transfer which is likely to determine crossing to the rest of the nearby receptor field and the whole of the distant field. Work in this second year will result in a statistically rigorous sampling scheme and efficient detection methodology which will be used in the third and fourth years to examine, over a wide range of environments, the effects on whole-field crossing of percentage male fertility, separation distance, pollen barriers and the regional topography and configuration of fields. Recommendations will be made for management options that will keep the percentage crossing below specified thresholds (e.g. 1%, 0.5%, 0.1%).