Records on 11,700 roast chicken body weights, and on 2,400 feed:weight gain ratio and chick production of 560 females selected in each of four generation of selection (1992-1995) were used to compare genetic parameter estimates of body weight (PC), feed:weight gain ratio (F:WG) and chick production(PP) obtained by Henderson Method III from one analysis that included all the records with the restricted maximum likelihood estimates of the same traits obtained from the following analyses: 1 - analysis including all PC, F:WG and PP; 2 - analysis of two generations including or not including PP; 3 - analysis of three generations including or not including PP; 4 - analysis of four generations including or not including PP. No remarkable differences between heritability estimates obtained by the two methods were observed. However the genetic correlation between pairs of traits showed different values. The values of genetic correlation between BW and PC and between F:WG and PC obtained by Henderson Method III suggested some selection bias, as a result of changes in gene frequencies. REML estimates were more stable and better reflected the biological relationship between the studied traits, important for selection criteria definition.
Roast chicken; estimate; genetic parameter; REML; Henderson
