Measurement of the soil fragment size distribution has been found to be related to erosion processes, water retention, temperature, porosity and other soil properties that can affect crop growth and productivity. However, there is little agreement about the procedure to follow to determine dry soil fragment size distributions using flat sieves. Our objective was to determine the effects of dry-sieving duration and loading on the resulting fragment size distributions of 2 soils [Maury silt loam (Typic Paleudalf) and McAfee clay loam (Typic Hapludalf)] under contrasting management practices (undisturbed sod and recently roto-tilled). After air drying, the samples were dry-sieved with a vertical vibratory sieve shaker and the mass of soil retained on each sieve in a nest of 13 sieves, with aperture sizes ranging from 16.00 to 0.85 mm, was determined. The ability to discriminate between undisturbed and tilled soil samples varied with the mass fraction of fragments examined. In general, there was a greater probability of discriminating between sod and tilled soils, with a lower coefficient of variation, when the sieving duration was at least 30 s. Loading the uppermost sieve up to 30% of its volume did not significantly affect the results, and allowed for processing, in a single operation, composite samples taken with a probe of 5.4 cm diameter and 10.0 cm length.