Assessment of phenotypic and genotypic stability for seed and lint cotton yields of some cotton cultivars

Fourteen field experiments were conducted during the two successive seasons of 2014 and 2015 at seven different locations from the Northern Delta of Egypt i.e., (Kafr El-Sheikh, El-Beheira and Domietta) to Middle and Southern the Delta of Egypt (El-Menoufia, Dakahlia, El-Gharbia and Sharkiea), to evaluate eight Egyptian cottons included two long stable genotypes; Giza 86 and Giza 94 and six extra – long staple; Giza 45, Giza 87, Giza 88, Giza 92, Giza 93 and Giza 96. Analysis of variance for randomized complete block design with four replications was done for each location. Then combined analyses of variance were calculated for eight cultivars, seven locations over two growing seasons . The statistical analysis for phenotypic and genotypic stability was carried. Concerning Giza 86, Giza 87, Giza 88, Giza 92 and Giza 93 cultivars which are considered the most desired cultivars occupy the most areas cultivated by high production of seed cotton and lint cotton yields. However, Giza 96 had the widest range of environmental index for seed cotton yield and lint cotton yield. While Giza 45 had the closest one for seed cotton yield and lint cotton yield. The environments were the most important source of variation explaining 91.92% and 90.92% of the variance for seed cotton yield and lint cotton yield, respectively, followed by the cultivars which explained 5.37% and 6.35% from the source of variation for seed cotton and lint cotton yields, respectively and the interaction between the cultivars and environment represented 2.71% and 2.73% from the source of variation for the two traits, respectively. With respect to the two cultivars, Giza 87 and Giza 88 their bi values do not significantly differ from the unity (bi = 1) and had deviation from regression ( S 2 di) not significantly differ from zero and their lint cotton yields exceeded the average overall genotypes, which indicated average stability and relative adaptability of the cultivars pointed out. The great variation of the cultivars to the estimated λi statistics suggested that the relatively unpredictable components (the deviation from the linear response) of the cultivar x environment interaction variance may be more important than the relatively predictable component (the coefficient of linear response). Results illustrated that all studied cultivars for both seed cotton and lint cotton yields are sensitive to environmental changes and these cultivars are expected to give high yields either for seed cotton or lint cotton under favorable environmental conditions.