COMPARISON OF NONLINEAR MODELS FOR DESCRIBING THE GROWTH OF LOCAL CHICKENS RAISED IN TOGO

Abstract

Growth is a trait of economic importance in chicken genetic improvement that is often described by mathematical models having biologically significant parameters. The local chicken,commonly raised in rural areas of Togo, is generally described as a slow-growing chicken, but little is known about his growth parameters. Therefore, this study was designed to determine the nonlinear model that best describes the growth curve of local chickens. The Gompertz, Logistic and Richards growth models were used to fit growth weights recorded from 126 males and 150 female chickens using GCFIT program of SIMFIT package (version 8.1.4).Coefficient of determination (R2), mean square error (MSE), Akaikes information criterion (AIC) and Bayesian information criterion (BIC) were used as criteria for models comparison. By estimating the highest value of R² and the lowest values of MSE, AIC and BIC, Gompertz model appeared to be the best-fitting model for local chickensgrowth. The estimated asymptotic weight was 1992 g for males and 1367 g for females. Thus, male and female chickens weighed 732.86g and 502.98g and aged 10.57 weeks and 9.77 weeks respectively at inflection point. Maturation rates of the chicks were similar (0.139 week-1 for males and 0.141 week-1for females). The growth parameters of this model may therefore be useful in defining selection criteria for local chickens in Togo.