Correlation among yield, phenotipic stability and simultaneus selection in cotton (Gossypium hirsutum L.)

Abstract

Yield and phenotypic stability are necessary to release new cultivars. The aim of this study was to estimate the linear and rank correlation between fiber yield (RENFI), six stability indices: S2 i (Roemer), σ2 va (Plaisted and Peterson), bi and S2 di (Eberhart and Russell), Wi (Wricke), σ2 i (Shukla), and four simultaneous selection indexes by RENFI and stability: Pi (Lin and Binns), ER and SH (Eskridge) and Si (1) (Huehn) in 10 cotton genotypes evaluated through six environments in humid Colombian Caribbean. A randomized complete block design was used for each environment. The combined variance analysis showed highly significant differences for environments and GxA interaction. RENFI did not correlate linearly with stability indices, whereas between selection indices only SH presented a significant association level (r>-0.43), except SH (r=0.65*). The six stability indices showed positive and significant linear correlation between all (r≥0.71*) except bi , which only was correlated with S2 i (r=0.96**). RENFI only showed rank correlation (rs =0.76**) with S2 i . The σ2 va index showed significant and positive rank correlation (rs >0.68) with S2 di , σ2 i , Wi , ER and SH. The ranking for stability of genotypes was equal and perfect (rs =1.0**) when used σ2 va, S2 di , σ2 i and Wi . The rank correlation between simultaneous selection indices was only similar (rs >0.82**) between ER, SH and Pi. The simultaneous selection indices ER and SH favored selection, preferentially, for stability given its rs >0.66 significant and direct with σ2 va, S2 di , σ2 i and Wi .

Keywords

• adaptability
• environment x genotype interaction
• parametric and nonparametric methods