A COMPARISON OF NON-LINEAR MODELS FOR DESCRIBING GROWTH IN N'DAMA CATTLE

Abstract

Five non-linear models were fitted to growth data for female N'Dama cattle kept at the University of Science and Technology research farm, Kumasi, Ghana. Comparisons for goodness of fit provided an R2 of approximately 96%. The Bertalanffy model was preferred because of its consistency in predicting weights at all ages, while the Richards model was more difficult to fit due to problems of non-convergence. The Brody's model over-estimated weights at all ages while the Logistic model converged too rapidly, thereby underestimating mature weights. Correlations established between estimated parameters and derived statistics indicated a tendency for early maturing animals to grow to smaller mature weights while large estimates of mature weights were associated with late maturing intervals.