In a glasshouse experiment it was examined whether narrow grading and selection from a commercial soybean seed lot cultivar ‘IAS-5’, could improve the uniformity of the seed crop grown from it and thereby enhance yield, quality and uniformity of seeds produced. The classes created were: Control (original seed lot); Size-graded seeds (projected area measured by image analysis 37–46 mm2); Non-cracked seeds; Yellow seeds; Size-graded sound seeds (size-graded, non-cracked, yellow, non-wrinkled, non-etched). Compared to the control, percentage of emergence, survival and number of yielding plants were enhanced in crops from non-cracked, yellow or size-graded sound seeds. Differences in plant numbers did not result in differences in crop yield. The different seed lots also did not differ in crop uniformity: time interval between stages of plant development, plant height 20 days after sowing, yield components, physical or physiological quality attributes of seeds produced, and respective coefficients of variation were similar. Fewer plants survived in crops showing a larger variation in plant height 20 days after sowing, thus reducing differences in initial plant-to-plant variation. Creating more uniform crops by additional grading or selection of commercial seed lots may therefore not be promising.