Vol. 2, Issue 6 (2017)
Stability analysis using various parametric and non-parametric methods in single and three-way cross hybrids of maize (Zea mays L.)
Author(s): Ravindra Babu S, Gokulakrishnan J, Ramakrishna S
Abstract: Farmers are always on the lookout for stable and high yielding maize hybrids in the competitive seed market. G x E interaction is the main challenge faced by the maize breeders in the development of single and three way cross hybrids in private and public breeding programs. Looking to the several methodologies applied for stability, it was thought to apply five parametric stability procedures viz., Eberhart and Russel’s regression technique, Shukla’s stability variance (Sh-σ2), genotype grouping technique based on Francis and Kanenberg’s coefficient of variation (CVi), Lin and Binn’s Cultivar superiority index (Pi) and Harnandez’s Desirability index (Di) and three non-parametric procedures viz., Huhn’s rank stability measures Si1 and Si2 and Kang’s Rank Sum (RS) for assessment of 20 single and seven three-way cross hybrids evaluated at six locations, which represent important maize growing ecologies in India, during winter season of 2015. Spearman rank correlation coefficients were worked out to assess the association between stability procedures. All the three components in Eberhart and Russel model i.e., environment (Lin.), G x E (Lin.) and pooled deviations were found significant. Hybrid SC18 was found high yielding, stable and adaptable to all the environments. Hybrids SC9, SC10 and SC14 were found stable and well adaptable to favourable environments whereas hybrids SC16 and TWC24 were found not only high yielding and stable but also suitable for marginal environments and they would be preferred by farmers as most of them cultivate maize under marginal environments. Almost all parametric and non-parametric methods unanimously selected SC16, SC10 and SC18, in that order, as potential hybrids among single crosses and TWC24 among three-way crosses. Significant positive rank correlation coefficients were obtained between all possible pairs of S2di, Sh-σ2, CVi, Si1 and Si2 and Kang’s RS implying that anyone of these parameters could be used to describe genotypic stability. Significant positive correlation between parametric and non-parametric stability measures indicated that rank-based non-parametric methods can be used as viable alternative to the existing parameters. Single and three-way cross hybrids were equally stable and desirable as indicated by various stability measurers. Pi, Di and Kang’s RS emphasized SC hybrids as most high yielding and stable whereas Sh-σ2, Huhn’s Si1 and Si2 indicated that TWC hybrids are stable and assured yielders under marginal environments.