IMAGE-ANALYSIS OF WHOLE GRAINS TO SCREEN FOR FLOUR-MILLING YIELD IN WHEAT BREEDING

Image analysis of whole-grain samples was used to predict milling qual ity in wheat breeding to select for this aspect of quality, while pre serving the seed intact for sowing. About 66% of the variation in flou r yield for 38 grain samples could be explained by four factors comput ed from the images of 100 grains for each sample (mean of grain area, lengths of minor and major axes, and ellipsoidal volume), plus test we ight. Test weight alone accounted for only 17% of the variation. The s et of grain samples consisted of eight genotypes (three cultivars and five breeders' lines) grown at up to six sites. The method devised is suitable for a breeding program, being relatively low in labor require ment, not requiring time consuming positioning of the grains, and havi ng low cost (less than $3,000 plus personal computer and software). Th e results of this preliminary study should provide direction for furth er development of noninvasive analysis of milling quality.