Genetic variability, association between agronomic and morphological traits and seed cotton yield as well as its components were studied in 56 glandless cotton varieties using a rectangular lattice design with three replications. These varieties were planted in the field, Cotton Research Institute of Varamin in 1999. Stepwise regression showed that, seed number / boll, 50% span length, plant height, lint strength, seed index and lint resistance could explain variations in yield (R2=0.4774). Path analysis showed that internode length, boll weigth, plant height and sympodia number had the highest direct effects on yield/ plant. Due to their high direct and or indirect effects on yield, internode length, boll weight and boll/plant can be used as selection criteria to identifY the high yielding genotypes. Factor analysis introduced 6 factors among 30 variables which explain 63.2% of total variance and included seed characteristics, plant morphological characteristics, yield and its components, maturity, lint length and lint yield. Cluster analysis using Ward method, based on Euclidean distance of 8.5 grouped the varieties into five clusters. The varieties in group 4, from the viewpoint of the height of 151 sympodia, monopodia number and node number to 1st sympodia, was the earliest maturity group and second group because of having high quantitative and qualitative seed and lint yield was distinguished as the best group.
Key words: , , , Multivariate statistical
methods.
)