evaluation of models to describe temporal growth in local chickens of ghana

the logistic, gompertz, richards and asymmetric logistic growth curve models were fitted to body weight data of local ghanaian chickens and french sasso t44 chickens. all four growth models provided good fit for each sex by genotype growth data with r2 values ranging from 86.7% to 96.7%. the rate constant parameter, k, ranged between 0.137 and 0.271 and were significantly different from zero for all genotype by sex groups. predicted mature weight from the four models ranged from 2840 g to 3020 g for sasso t44 female, 3225 g to 3448 g for sasso t44 male, 1170 g to 1332 g for ghanaian female and 1607 g to 1777 g for ghanaian male chickens. for the richards and asymmetric logistic functions, the shape parameter (n) which influences the point of inflection ranged from -0.126 to 0.713, indicating that the shape of each of the genotype by sex sigmoid function is negatively asymmetric. between the two simpler models, with fixed inflection point, the logistic function was characterised by a younger age at start of the growth acceleration phase, older age at the point of inflection, younger age at the end of the growth deceleration phase and lower mature body weights relative to the gompertz function. based on the bayesian information criterion (bic), the gompertz function was preferred to the logistic function. the r2, bic values and predicted body weights for the asymmetric logistic function were similar to those of the richards function. these complex models with flexible inflection point provided better goodness of fit relative to the gompertz model. therefore it is concluded that, where data structure and availability of adequate computing power permit, models with flexible inflection point such as the richards function can be used to provide accurate parameter estimates for the characterization of growth of indigenous chickens.