Univariate stability analysis of genotype×environment interaction of oilseed rape seed yield

Abstract

Thirteen stability statistics were used to analyze genotype × environment (GE) interaction of 36 canola genotypes. Combined analysis of variance indicated that GE interaction significantly influenced seed yield performance. According to Type I stability concept (environmental variance, coefficient of variation and stability variance) genotypes G7, G9 and G13 were the most stable genotypes, while based on the Type II concept (coefficients of three linear regres sion models), genotypes G33, G27 and G29 could be selected as the most favorable genotypes. Also, genotype G7 was the most favorable genotype according to Type III stability concept (deviation from linear regression method). Genotypes clustering based on stability properties and mean yield grouped them into three distinct classes. Coefficient of determination for the canola genotypes indicated that genotypes G27 and G33 were the most stable genotypes but the genotypes G1, G10 and G25 had the highest desirability index and were the most stable ones. The plot of principal component analysis was used for graphic display of the relationships among statistics and the first axis distinguished the Type II of stability concept from other types and mean yield groups near this stability type. However, based on most statistics and mean yield performance, genotypes G9 or Fanaei-6 (2592.47 kg ha-1), G11 or Fanaei-14 (2592.47 kg ha-1), G12 Fanaei-15 or (2592.47 kg ha-1) and G19 or Dez-7169 (2592.47 kg ha-1) were the most stable and favorable genotypes and are recommended for national release Iran.