We report a study carried out to compare the sphericity index values calculated on pelleted seed samples using three different measurement principles:  a commercially-available image analysis method (3D) that captures views of individual seeds falling under gravity;  two different (2D) image analysis method systems that captures images of individual seeds lying flat on a flat surface;  size screening of a population of seeds using round-hole and slot-hole sieve screens. To calculate the sphericity index (value = 1 for perfect sphericity): in  and  ratios of minimum and maximum diameters determined by computer algorithms were used to calculate a sphericity index for individual seeds; in  a ratio is calculated between the median seed-size values determined by the two series of screens. Pelleted seed with three basic different geometric ellipsoid shapes were used, made from seed of brassica, lettuce and tomato; the 3 shapes were mixed in different proportions to produce six ‘unknown’ samples for comparative analysis. Results showed close conformance between the sphericity values obtained by the three laboratories equipped with the same 3D system. Results from the 2D systems produced higher means sphericity index values than the 3D system, and also differed from each other. The size screening analysis in general produced sphericity values intermediate between the 3D and the two 2D systems. These results can be partly explained by the different geometric aspects of the seed measured by the three systems.