Grain yield stability estimates of late maturity maize hybrids

Abstract

With global climate change including unpredictable geographic and temporal weather patterns that lead to significant genotype ? environment interaction (GEI) the maize performance assessment would need to be complemented with stability analysis. The objectives of this study were: i) estimation of parametric and non-parametric stability indices for grain yield of late maturity maize hybrids, ii) assessing correlations and grouping of stability indices for grain yield of late maturity maize hybrids. The eight maize hybrids of FAO 700 maturity group (ZP1-ZP7, and check Ch) were evaluated for grain yield in the multi-environment trial including five locations in Serbia during 2020. and 2021. year. Stability analyses included twenty parametric and non-parametric indices, Principal Component Analysis (PCA) and Pearson?s correlations between stability indices. The sum of ranks for tested late maturity maize hybrids were in the interval from 37 (ZP6) to 146 (ZP4). The average rank was in the range from 1.8 (ZP6) to 7.3 (ZP4). The interval of variation for standard deviation of average ranks for tested maize hybrids was from 1.3 (ZP4, Ch) to 2 (ZP2) indicating satisfying accordance of utilized parametric and non-parametric stability measures. The highest average grain yield across all tested environments was observed for ZP1, which was second most stable hybrid. The most stable late maturity maize hybrid tested over ten environments was proven to be ZP6 with sum of ranks and average rank of 37 and 1.8, respectively, with standard deviation of average rank 1.5. PCA biplot showed two groups of parametric and non-parametric stability estimates: I group-coefficient of variation (CVi), Eberhart and Russel?s regression coefficient (bi), Perkins and Jinks?s regression coefficient (Bi), GEI variance component (?(i)), coefficient of determination (Ri2); II group-Eberhart and Russel?s deviation from regression (S2di) Wricke?s ecovalence (Wi2), Shukla?s stability variance (?i2)Perkins and Jink?s deviation from regression(?i2), superiority measure (Pi), the mean of the absolute rank differences of a genotype over all tested environments (Si(1))the variance among the genotype ranks over all tested environments (Si(2)), the sum of the absolute deviations for each genotype relative to the mean of ranks(Si(3)), the sum of squares of rank for each genotype relative to the mean of ranks (Si(6)), NPi(1), NPi(2), NPi(3), NPi(4)- Thennarasu non-parametric stability estimates, mean variance component (?i), Kang? rank-sum (KRi). Pearson?s correlation bring along the redundant stability indices and one can choose to use Wi2 or ?i2, bi or Bi, Wi2or ?i, ?i2 or ?i.