Estimation of Heritability, Genes Number and Multivariate Analysis Using Non- segregation and Segregation Generations in Two Cotton Crosses

Abstract

This study was conducted to evaluate the heritability methods, the genes number equations and comparison between them, as well as multivariate analysis of cotton yield and its components traits in the two crosses G.85 x TNB and G.86 x Suvin. Broad sense heritability (BSH) using Mahmud & Kramer (1951), Burton (1951), Weber & Moorthy (1952), Modified Weber & Moorthy (1952), Briggs & Krowles (1967), Mather & Jinks (1971), Lawrence & Jinks (1973) and Kotecha & Zimmerman (1978) methods, as well as narrow sense heritability using Warner (1952) and Modified Warner (1952) methods were calculated. The methods of BSH and NSH showed high values (BSH < 0.60) and significant for yield and yield components traits in the two crosses. For BSH estimations, the highest values by Mahmood & Kramer and Burton methods and the lowest values by Mod.Weber & Moorthy were registered for most studied traits. Estimates of genes number affecting traits were obtained with Chen & Line (1995). The genes number values by equation N3 were much higher than the other equations (close to each other or slightly different) for all studied traits in the two crosses. Based on the ranks method and cluster analysis suggested that there are differences between most the methods of BSH and genes number estimations. The methods of Mather & Jinks and Lawrence & Jinks gave the same values, the two methods of Mahmud & Kramer and Briggs & Krowles as well as the two methods Weber & Moorthy and Kotecha & Zimmerman showed equal or close values for the studied traits in the two crosses. While, the others methods of BSH had difference with these methods and with from each other. These methods are calculated based on the components of variance, so a change in each component of the variance can affect it. Thus, these methods differ with respect to the calculation of an environmental variance. According to principal component analysis (PC), the PC1 and PC2 had mainly distinguished the generations in different groups. The PC1 and PC2 contributed towards 86.29% and 92.36% of cumulative variability in the two crosses G.85 x TNB and G.86 x Suvin, respectively, and the PC1 exhibited Eigen value >1 for all studied traits in the six populations. According to biplot and based on the all populations, the PCs with the highest variability showed positive correlation to yield and its components, but, they differed in their degree of significance/insignificant and consistency in quantity. The PC of the relationship between the six generations revealed that the most appropriate generations for selecting these traits were BC1 generation in the cross G.85 x TNB as well as F2 and BC1 generations in the cross G.86 x Suvin. Backcrossing may be done for 2–5 cycles (BC2 – BC5) at Suvin parent for improving cotton yield in Egypt.