The aim of this work was to study the adjustment of Richards, Brody, Gompertz, Von Bertalanffy and Logistic functions on the growth curve of Santa Inês hair sheep and to estimate genetic parameters for the traits obtained from the best fitting function. Information from females of the Embrapa Tabuleiros Costeiros herd recorded between 1993 and 2004 and from the Embrapa Caprinos herd recorded between 1981 and 2004 was used. Functions were fitted for each herd, using NLIN procedure of Statistical Analysis System software (SAS), by GAUSS method. The coefficient of determination (R²), residual mean square (QMR) and mean prediction error (EM) were the criteria used to determine the best fitting function. For the Embrapa Tabuleiros Costeiros herd, all the functions subestimated the weights, except the Richards function. On the other hand, all the functions superestimated the weights of animals from the Embrapa Caprinos herd. The best fitting in the both herds was obtained using the Richards function. The respective values of mature weight and maturation rate estimated by Richards function were 54.38 kg and 0.00144/day for Embrapa Tabuleiros Costeiros herd and 42.74 kg and 0.00260/day for Embrapa Caprinos herd. The Richards function was used to estimate individual growth curves of animals in order to estimate genetic parameters and the (co)variance components for traits of economic importance using the Derivative Free Restricted Maximum Likelihood method (DFREML). The estimates of direct and maternal heritabilities ranged, respectively, from 0.01 to 0.99 and 0.00 to 0.13 suggesting the possibility of changing the growth curve of Santa Ines sheep by selection.
animal model; heritability; mature weight; maturing rate; non linear function; Richards function
