Classification and Comparison of Eight Different Growth Curve Methods for Pekin Duck

As in all livestock species, growth is the most important trait. Growth is an increase in size (height, length, weight) with advancing age and growth curve models provide a visual assessment of growth as a function of time. These models can be used for predicting body weight for a specific age from a dimensional perspective.

In this study, we compared Brody, Gompertz, Logistic, Gamma, Schnute, Richards, Negative Exponential and von Bertalanffy models on the body weight of Pekin ducks raised in Türkiye using 109 female and 110 male birds for 10 weeks of age.

Ducks were reared with a feeding program identical to standard commercial practices. Growth models were fitted to the data of the average growth curve and for the individual growth curves. Parameters were estimated using the SAS 9.0 statistical package program, Proc Nlin procedure and Gauss-Newton algorithm were used to model the curves.

3-D clustering results showed that the models clustered in three clusters both female and male Pekin ducks. According to the goodness of fit criteria such as mean square prediction error, coefficient of determination, adjusted coefficient of determination, accuracy factor, bias factor, Durbin-Watson value, Akaike Information Criteria, corrected Akaike Information Criteria, Bayesian Information Criteria and standard error of the regression and interpretation of the growth curves and clustering hierarchical 3-D dendrograms, the Schnute and Richards models, were found to be a suggestible model for 70 days growth of both female and male Pekin ducks, but the Richards model can be more recommendable due to calculation easiness.