The estimated genetic parameters were used for selection of important conformation traits and reduction in the number of the described characteristics. For reduction in the number of described traits from the original 36 to 24, the traits were selected according to: measure of genetic similarity (cluster analysis), measure of uncertainty multidimensional quantity, value of the variance of aggregate genotype, value of the variance of selection index and correlation of trait to the first principal component of the genetic matrix. Reduction in the number of the described traits was based on the value of heritability coefficient and genetic correlations matrix. The reliabilities of selection indexes were estimated between 0.41 and 0.53. Among the three multivariate analysis methods evaluated in this study, the variance of selection index had the highest reliabilities of selection indexes. The estimation of selection index variance which omitted traits with low heritability coefficient and high genetic correlation was the most suitable for the traits selection. This procedure would enable the breeders to reduce field costs (e.g. time, labour) required for obtaining the genetic parameter estimates necessary for a specific breeding programme.