Publicación:

Growth curves modeling of multirracial replacement milk heifers reared under pasture tropical conditions.

Resumen

Nonlinear mixed models were used to describe the growth pattern of dairy replacement heifers from birth to conception age using the Gompertz, Von Bertalanffy, Verhulst and Logistic functions. The studied racial groups were Holstein-50% x Gyr-50% (HG), Holstein-50% x Jersey-50% (HJ) and Jersey-50% x Holstein-25% x Gyr-25% (JHG). Random effects were linked to asymptotic body weight (β0), constant of integration (β1) and maturation rate (β2) parameters in each mathematical function. Models with fixed effects and different combinations of random effects were compared according with their capacity to fit the data set. Residual variance, Akaike and Bayesian information criterion and residual plots were used to evaluate the models. In general, the addition of random effects in the models increased the accuracy in parameters estimates, however the best fit was achieved with the Von Bertalanffy function with random effects associated to β0, β1 and β2 parameters. This model improved the likelihood statistics and reduced the residual variance in 47% on average with respect to the fixed effects models. Compared with an optimal growth process under tropical condition, model predictions indicate that heifer within the three evaluated genotypes are reaching conception with inadequate weights as a consequence of the poor diet quality available during the rearing period. The results suggest that exist the opportunity to exploit the real potential growth of heifers improving the rearing management, nutrition and selecting for precocious animals.