Previous studies of the possible bias of falling-stream sample cutters have used physical experiments, two-dimensional and three-dimensional DEM (discrete element modelling) models with spherical particles. The present study uses super-quadric particles describing particles with variable aspect ratios and variable blockiness, and is more realistic than previous modelling studies. Our results support the commonly used rule that the cutter aperture should be at least three times the nominal top size of the material being sampled, but do not support a commonly used rule relating acceptable cutter aperture to cutter speed and suggest that particle shape, belt load and distance of cutter below the head pulley are also key factors which can influence cutter bias. (C) 2008 Elsevier Ltd. All rights reserved.