This paper examines the success of selection for preharvest sprouting tolerance in white-grained bread wheat using a standard wetting treatment, germination of hand-threshed seed and falling number measurements. The rain simulator was usefull in shifting the population mean of field grown material towards higher levels of tolerance in successive years; however, large genotype x year interactions in material sown under rain protection did not allow accurate assessment of individual genotypes. The most accurate assessments were achieved using falling number measurements (h2 = 80.7%) and hand-threshed seed germinations (h2 = 38.4%), where no genotype x year interactions were recorded. Seed dormancy (determined from hand threshed grain) correlated significantly with change in falling number following 3 days' treatment in the rain simulator ( r = -0-56**). Visual measurements scored in the rain simulator, however, did not correlate significantly with seed dormancy in the first year (r = 0.20) but correlated strongly in the second (r = 0.73***). In comparisons of the same test between years, falling number (without rain treatment) and seed dormancy were significantly correlated (r = 0.68* and 0.90***, respectively), whilst visual scores of sprouting showed no association (r = -0.03).