UNIVERSIDADE DE TRÁS-OS-MONTES E ALTO DOURO

vi In a first stage, we used, in a Bayesian approach, a threshold-linear mixed animal model, which combine PROL and PL120. The first trait was associated with a latent continuous variable through data augmentation. In a second analysis, we implemented a hierarchical Bayesian scheme with non-linear Wood function modelling lactation curve. The estimated posterior means of the genetic parameters for the White/Black strains were, respectively, to the PROL and PL120 variables: 0.06/0.07 and 0.12/0.14 for heritability; 0.09/0.10 and 0.14/0.18 for repeatability; 0.44/0.28 for the additive genetic correlations; 0.83/0.85 for the permanent environmental correlations and –0.01/-0.02 for the residual correlations. The genetic and non-genetic parameters for the White/Black strains, concerning the Wood parameters (A, B and C) were, respectively: 0.09/0.11; 0.21/0.24 and 0.17/0.14 for heritability; and 0.10/0.12; 0.39/0.40 and 0.19/0.17 for repeatability. Genetic correlations between parameters are intermediary to high, with the exception for B-C, in the White strain that is almost zero. The residual correlations are middle values (∼0.55) between A and B parameters, high (∼0.90) for A-C par and lower (0.14 to 0.22) between B and C. From the results of this study, we can conclude that is possible to achieve genetic progress for milk yield with positive genetic correlation in prolificacy. By the other hand, in spite of the heritabilities of curve parameters have low mean values, the genetic capacity of changing milk production curve is assured, also by the genetic positive correlations between Wood parameters.