Knowledge of pollen dispersal is essential for maximizing cross-fertilization in apples (Malus ×domestica Borkh.) and achieving optimal orchard design. Using allozyme markers, we examined dispersal of pollen from trees of a single cultivar (`Idared') throughout two apple orchards. In each orchard, the percentage of seeds sired by `Idared' was estimated for trees sampled at regular intervals along three transects, extending up to 18 rows (86 m) from the closest donor trees. The percentage of seed sired by `Idared' pollen ranged from 76% to 1% of seed sampled for a row. No differences in pollen dispersal were found among transects, despite differences in proximity to the bee colonies. Variation in `Idared' siring success was attributable to the cultivar of the fruit-bearing trees as well as their distance to the nearest `Idared' tree. Cultivar effects were associated with differences in flowering overlap, but not cross-compatibility with the pollenizer. Furthermore, flowering overlap was a good predictor of siring success only when the flowering times of competing pollenizer cultivars were also considered. The implications for orchard design are discussed.