A computer simulation experiment compares the parametric stereological method of Barbery, as extended by Leigh, Lyman and Gottlieb, to the nonparametric method developed by Gay. The simulation embraces random particle shapes, and textures generated using random processes. Special deterministic textures are included in order to pose difficult stereological problems. Both methods are found to perform well in cases likely to occur in mineral processing. In the difficult cases, Gay’s method remains accurate. Both methods perform well on samples containing 2000 particle sections, and moderately well on samples of only 250 sections.