Grain yield stability of corn hybrids using best linear unbiased predictors

摘要

Mixed model methods provide best linear unbiased predictors (BLUP) that can be used to quantify location-specific effects and genotype by environment interaction effects. In this study, data from multilocation corn yield trials were used to obtain BLUP for corn hybrids (Zea mays), the main effects of their inbred parents and their interactions with locations to determine stability of performance of hybrids and their parents across environments. Thirty-two female lines and 16 males combined in 39 hybrids, plus one commercial control, were evaluated across 22 locations of the U.S. Corn Belt during 1996. A combined analysis of data, using a mixed model with both genotype and location effects specified as random, was done. Genotype by environment interactions were significant for hybrids and female parents, but not for male parents. BLUP was as effective as Eberhart and Russell's stability parameters in detecting two stable and high performing hybrids across environments.