Over a 10-year period, runoff and soil erosion on the plots of the Woburn Erosion Reference Experiment were concentrated in periods with sparse vegetation cover: in winter after the late planting of cereals; in spring after the planting of beets; or when soils were bare after harvest. The mean event runoff of 1.32 mm from plots cultivated up-and-downslope was significantly greater (P<0.05) than that from plots cultivated across-slope (0.82 mm). However, mean event soil loss was not significantly different between the two cultivation directions. No significant differences were found between minimal and standard cultivations. Mean event runoff from the across-slope/minimal tillage treatment combination (0.58 mm) was significantly less (P<0.01) than from the up-and-downslope/minimal tillage (1.41 mm), up-and-downslope/standard tillage (1.24 mm), and across-slope/standard tillage (1.07 mm) treatment combinations. Runoff from the across-slope/standard treatment combination was significantly (P<0.05) less than from the up-and-downslope/minimal tillage treatment. The across-slope/minimal tillage treatment combination had a significantly smaller (P<0.05) event soil loss (67 kg ha−1) than the up-and-downslope/standard tillage (278 kg ha−1) and up-and-downslope/minimal tillage (245 kg ha−1) combinations. Crop yields were significantly (P<0.05) higher on across-slope plots in 1988, 1996 and 1997 than on up-and-downslope plots, and were also higher (but not significantly) on the across-slope plots in 7 of the 8 remaining years. Minimal cultivation decreased yield compared with standard cultivation in one year only. We recommend that across-slope cultivation combined with minimal tillage be investigated at field scale to assess its suitability for incorporation into UK farming systems.