Curva de crescimento de codornas de corte por meio de modelos de regressão não-lineares

Abstract

The aim of this study was to evaluate different non-linear regression models to describe the growth curve of male and female meat quail. The study involved 30,410 records of body weight originally measured weekly from the 1st to the 42nd day of life of meat quail participating in the Breeding Program of the Federal University of Vales do Jequitinhonha e Mucuri. Body weight was separated into the male and female classes and subjected to the non-linear Brody, Gompertz, Logistic, MMF (Morgan-Mercer-Flodin), Richards, and van Bertalanffy models. The models were evaluated with respect to the adjustment of predicted growth curves in relation to observed values. No convergence was obtained with the Brody model. The other models yielded good AI (asymptotic index) values, and the MMF model was the most suitable to predict the weight of males, with an AI of 33.59, whereas the Gompertz model, with an AI of 36.100, was the most suitable to predict the weight of females. Therefore, in the present study, the Gompertz and MMF models best fitted the data for males and females, with an R2aj greater than 0.92. Results point to the use of both methods to describe the growth of these animals, irrespective of sex.