The objective of this research project was to determine the potential for selection of maize (Zea maize L.) genotypes that have breakage-resistant grain. Eighty random inbred lines derived from BSSS and 40 single crosses developed from these inbreds were grown in 1976, 1977, 1978 and 1979 at two locations in each year. Data were collected in machine-harvested grain samples in each plot for harvest moisture, breakage percentage, kernel weight, kernel volume, specific gravity, and the Fast Green dye test. Data for visual rating and day to anthesis were collected from each plot;The analyses of variance revealed highly significant differences among both inbred and single-cross genotypes for all traits. Significant genotype x environment interactions were obtained for all traits, except specific gravity and days to anthesis. The relative size for the estimated genotype x environment variance component for each trait was from 16 to 58 percent as large as the estimated genotypic variance component. Progeny-mean heritability estimates were relatively large (76-87%), except for specific gravity (39%);Correlation analyses, for both the inbreds and single crosses, revealed strong interdependence among some plant characters, i.e., percentage breakage with harvest moisture, visual rating, kernel weight and volume, specific gravity, and the Fast Green dye test for both the inbreds and hybrids. The correlation between percentage breakage and the Fast Green dye test was not as high as was expected. Inbred-hybrid correlations revealed large trait-to-trait correlations, except for days to anthesis, which indicated that expression of a trait in inbred parents gave a fairly good indication of the expression of the same traits in their hybrid progenies. Selection against large seeded, floury kernel-type inbreds may be a useful screening device to eliminate the more breakage-susceptible material provided potential yield loss from smaller kernels could be controlled;The inbreds were ranked with various rank summation, Elston's weightfree, and restricted indices and the ranks correlated with Spearman rank correlations. The restricted indices were the most effective in selecting breakage-resistant genotypes because seed size loss could be controlled. The rank summation indices may be an alternative to the more complicated restricted indices. However, selection differentials revealed seed size losses resulting from selection with both the rank summation and the Elston indices that may deem restricted indices essential for breakage-resistance selection.