Genetic divergence among broiler lines (L1, L2, L3 and L4) was studied by multivariate analysis. Four economic important traits were evaluated in three periods (initial, medium and total): age at first egg, egg number, average egg weight and body weight. The study of the genetic divergence among lines included the following techniques:Multivariate analysis of variance, cluster analysis using Mahalanobis generalized distance and Tocher optimization method, and canonical variables analysis. Significant difference was obtained by the larget Roy eigenvalue test. The cluster analysis of traits in the initial (L1 and L2; L3 and L4) and medium (L2 and L4; L1 and L3) periods grouped the four genotypes in three groups, and in total period the genotypes were grouped in two groups (L2, L3, L4 and L1). Canonical variables analysis showed that the first two canonical variables accounted for more than 96% of the total variation in the three periods. This allowed to represent the performance of genotypes in a bidimensional dispersion graph. In the total period of production study the genetic divergence was lower.
Broiler; multivariate analyses; cluster analysis; canonical variables
