GENOTYPE × ENVIRONMENT INTERACTIONS FROM YIELD TESTS AND THEIR APPLICATION TO SORGHUM BREEDING PROGRAMS

Abstract
Estimates were obtained of genotype x environment interaction in yield tests of 21 grain sorghum hybrids at 7 locations for 2 years in Kansas. Genotype x location and genotype x year interactions were significant but their components were small compared with second-order interaction and the genetic variance. Upward bias of variance components of genetic and genotype x location interaction when years were analyzed separately indicates that results from a single year are inadequate. Comparisons of standard errors of genotype means for several combinations of replications, locations, and years suggested an optimum combination of 3 replications, 10 locations, and 2 years of testing.