A many-sided evaluation of grain yield stability of 22 maize hybrids from the early group (FAO 300-400), tested in 4 locations was done. Different methods and models were applied: variation and regression analysis; AMMI and cluster analysis as well.
A modified method for grouping of hybrids by mean yield and mean values of the regression coefficient bi was proposed. A good accordance between that method and AMMI-1 analysis was observed. The cluster analysis also confirms this consent.
Calculated theoretical yields by locations create conditions for more precise zoning of hybrids according to environments.