This work was accomplished with the purpose of establishing quantitative relationships between the D 50 index of the size distribution of the soil-eroded sediments, the runoff velocity, the SR index of the tillage-induced soil surface roughness, and the mean weight diameter (MWD) of the soil aggregates, in a soil submitted to different forms of management. The study was developed in the field, at the Agricultural Experimentation Station of the Federal University of Rio Grande do Sul (EEA/UFRGS), in Eldorado do Sul (RS), Brazil, by applying simulated rainfall on an Ultisol with a sandy clay loam texture in the surface layer and 0.115 m m -1 average slope steepness. This soil had been put into agricultural use by different manners (continuous and discontinued cultivation), with different crop sequences (winter and summer, grass and legume crop species, planted in rows, using no-tillage), for a 7.5 year period (starting at the original condition of native pasture). Seven erosion tests were performed in the study, each one of them at 63.5 mm h -1 rainfall intensity and 1.5 h duration, using the rotating-boom rainfall simulator and 3.5*11.0 m experimental plots. The referred erosion tests were performed in the following soil surface physical conditions: (a) non-mobilized soil, with complete and no cover by crop residues, and (b) soil successively mobilized by the passage of a light disc-harrow (five times, one at a time), with no cover. It was observed that the crop sequences provided values of the MWD index significantly different each other, which reflected in significantly different values of the SR index and, as consequence, of the runoff velocity and the D 50 index, with the sequences with none or less time of discontinued cultivation (in the last period of the research) having produced the best results. In the non-mobilized, completely mulch-covered soil, with a firm and smooth surface, the mulch of crop residues was the dominant factor either in reducing the runoff velocity or in trapping the eventually detached soil particles of larger size, which led to very small values of the D 50 index. In the non-mobilized, uncovered soil, where runoff reached its highest velocities, the size of the eroded sediments was determined by the consolidation of the soil surface and by the values of the MWD index, being the smaller for the greater values of the two last variables mentioned. Yet in the soil successively mobilized by the one at a time passage of a light disc-harrow and bare, with a loose and rough surface, the roughness of the soil surface created by tillage was the dominant factor either in reducing the runoff velocity or in trapping the detached soil particles of larger size, which increased the percentage of eroded sediments
