A weight-age growth curve was studied in 160 bovine Hereford males. It was fitted to the Brody, Gompertz, Logistic, Richards and von Bertalanffy functions. The parameters estimates for the nonlinear regression model were calculated by the Weighted-Least-Square through MODEL procedure (SAS/ETS, 1995). An animal individual adjustment was obtained by the two models: non-weighted and weighted by the inverse of variance. The comparison with the previous models was made by the parameters biological interpretations and by fit quality evaluators (Adjusted determination coefficient, error mean square, mean predicted error, Student's t-test for the parameters A and K estimates and convergency percent). The best fitted model was the weighted by the inverse of variance. The functions that best fitted the data were those of von Bertalanffy and Gompertz followed by Logistic. Brody and Richards functions presented a low convergency percent, being inadequate to describe the data in study.
nonlinear regression; heterogeneity of variance; growth curve
