ABSTRACT - The objective of this work was to evaluate seven non linear functions to describe Guzera male and female growth curves, and their crossbreds (3/4Guzera-1/4Brown Swiss, 1/2Nellore-1/4Brown Swiss-1/4Guzera, 1/2Guzera-1/2Brown Swiss, 1/2Chianina- 1/4Brown Swiss-1/4Guzera, 1/2Caracu-1/2Guzera, 1/2Caracu-1/4Brown Swiss-1/4Guzera) submitted to three levels of supplementation (without, during the dry period and the whole year), to analize and to select one function which presented lower regression deviation, lower residual sum of squares, higher coefficient of determination, lower variation among animal parameters, higher convergence and assintotic compatible value with the observed mature weight. The function that best express these parameters of selection was the Gompertz, followed by the Von Bertalanffy function. The Richards function showed the high computational difficulty to the adjustment, resulting on lower convergence. Brody functions presented the largest and the modified Gompertz and Logistic functions showed the smallest mature weight.
beef cattle; nonlinear functions
