Non linear alternat

Abstract

The objective was to examine alternative non linear models to describe grouw in goat breed Mambrino. Data were used to weigh 18 (monthly) from 44 animals frombirth to 540 days of age. For the selection of models, data were generated by the evaluators AIC, BIC, MEP and R_2aj, R_2p and performed a cluster analysis. The  non linear models were used Schute, Mitscherlich, Meloun I, Meloun II, Meloun III, Meloun IV, Reynolds, Gamito, Meloun V, Meloun VI, Michaelis – Menten, Michaelis – Menten Modified and Weibull. Of the thirteen models used three had no convergence (Reynolds, Michaelis-Menten  Modified and Weibull) and four models showed no conventional pattern of growth curves (Schnute, Meloun II,   Meloun V and Gamito). The Michaelis-Menten model was the best in describing the growth of animals followed the models Mitscherlich  and Meloun III.  The gender did not significantly influence the shape parameters of the model Michaelis-Menten  and observed a significant reduction in animal growth after weaning, especially after 165 days, which suggests specific food handling measures in this phase.