Cool-period soil erosion from farmland in Prince Edward Island occurs predominantly through rilling mainly due to snowmelt. This is mostly evident on fields coming out of potatoes (Solanum tuberosum L.), a crop that is traditionally cultivated as part of a 2-yr rotation with cereal grains, or a 3-yr rotation with cereal grains and a forage crop for hay. Accurate assessment of snowmelt erosion (associated processes and effects) has been historically constrained by data inadequacy, particularly for rilling; and although a number of models have now been developed for estimating cool-period erosion, they still stand to benefit from the generation of actual field data. In the present study, actual volume measurements were made of erosion forms, for about 80 site-years, after the spring snowmelt in fields at various stages of potato rotation. Cool-period erosion, predominantly rills, amounted to about 30 t ha−1 for 2-yr and 3-yr rotations, thus an annual seasonal average of 15 t ha−1 and 10 t ha−1, respectively. Potato fields left untilled over winter averaged about 20 t ha–1 which increased to 36 t ha−1 with fall ploughing. Sediment deposition in fields coming out of potatoes was 13.3 t ha−1 compared with the amount of 0.4 t ha−1 in fields coming out of forages and 1.6 t ha−1 in fields coming out of grain. Fields going into the winter in sod or stubble showed negligible erosion, averaging about 0.1 t ha−1 in spring. The results of this study demonstrate the dominant influence of cropping practice in the erosion process during the cool period, and the tendency for traditional computational procedures to underestimate soil loss. Key words: Cool-period erosion, rilling, rill volume, sediment deposition, soil-loss prediction, erosion mapping, crop rotation