The objective of this paper was to compare different random regression models adjusted through Legendre polynomial functions of different orders to evaluate the one best adjusted to the genetic study of the meat quail growth curve. The data used in this study were from 26,835 and 27,447 observations from 3,909 and 4,040 meat quails of the UFV-1 and UFV-2 strain, respectively. The body weight of the two strain was evaluated on 1, 7, 14, 21, 28, 35 and 42 days of age. The orders of adjustment of the continuous functions were gradually increased (orders varying from 3 to 6), to determine the least necessary order to describe structure of covariance in function of time in the random regression models. Legendre polynomial functions of the 6th order for direct additive genetic effect and 5th order for permanent effect, for strain UFV-1 and 6 for both random effects for strain UFV-2 should be used in the genetic evaluation of the quail meat growth curve in study. The estimated heritability indicates which genetic traits can be obtained as a response to the physical body weight selection in two strain.
heritability; Legendre polynomial; variance component
