The knowledge of the genetic relationships between the parameters of lactation curve and the milk yield is important for the selection of cows and sire. The records milk yield from 583 cows of the Guzerá breed, daughters of 165 sire, were used to estimate the lactation curve parameters by two mathematical models. Milk productions were obtained from the extracted base of data from National Animal Science Archive maintained by National Center for Dairy Research (Embrapa Gado de Leite). The analyses were performed by the MTDFREML system, under animal model containing herd-year-season of calving, the linear and quadratic age effects of calving, as fixed effects animal, permanent environment and error included as random effects. The estimates of genetic correlations between the total milk yield and the initial milk yield, and the rate of decline of milk yield were 0.55 and 0.86, and -0.19 and 0.38, respectively, for each mathematical model. For last two characteristics, these estimates were 0.45 and 0.75, indicating that cows with larger initial milk yield presented accentuated milk yield declines along the lactation period. The genetic correlations among 305-days yield and the parameters of model followed the same tendency, in magnitude and sign, to the one of the total milk yield. In agreement with the results, it can be inferred that the selection for the total milk yield results in increment in the initial milk yield, and in inexpressive changes in the rate of decline of the milk yield, not altering the form of the lactation curve. The rank correlations among the classifications of the predicted genetic values were high and positive (0.68 and 0.74), suggesting that the individuals with the highest genetic values for initial milk yield, would be also those with higher values for total milk yield.
genetic correlations; Guzera breed; lactation curve; mathematical model; REML
