Grouping pattern of maize test locations and its impact on hybrid zoning

Abstract

Grain yield is a result of the combined effects of genotype, environment and their interaction. The importance of the genotype x environment interaction is that it causes different reactions of maize hybrids when they are grown in different environments. This study was performed to analyse discriminativeness, representativeness and relationships of test locations for a grain yield of nine maize hybrids grown at nine locations in a 5 year period 2007-2011 in the South Pannonian Basin, a part of central and south European corn belt. A three-way mixed model revealed the highly significant (P lt 0.01) REML variance components for year x location, year x location x hybrid and residuals. The grain yield was also significantly (P lt 0.05) affected by the year x hybrid and location x hybrid interaction effect. The sites regression (SREG) model was used to determine whether mega-environments exist, and/or whether the test locations used are suitable for maize hybrids zoning. To determine which environmental factors mostly affect the hybrid x location interaction, the partial least squares regression approach was applied. The "which-won-where" pattern of the SREG biplot confirmed a rank change interaction between the locations, indicating the presence of strong and unpredictable rank-change location-by-year interactions. The main characteristic of the grouping pattern was poor repeatability, since most of the locations that made one mega-environment tended to change from year to year. Therefore, it is preferable to develop hybrids with high yielding potentials, wide adaptability and stability, rather than to develop hybrids specially designed for narrow agro-ecological regions.