Growth patterns in broilers
DOI:
https://doi.org/10.1590/S1678-3921.pab1984.v19.16029Keywords:
growth curves, mathematical models, body weight, age, maturing rateAbstract
Polynomial regressions and nonlinear (Gompertz, Richards, Logistic and Bertalanffy) were fitted to age-weight data of male and female broilers from birth to 68 days. The experiment was carried out in Concórdia, SC, Brazil, from August to October 1981. The rations, supplied ad libitum, were isocaloric and isoproteic containing 2,970 kcal/kg of metabolizable energy (ME) and 22.8% of crude protein (CP) in the initial phase (first to 28th days) and 3,050 kcal/kg of EM and 19.8% CP in the final phase (29 to 68 days). The Gompertz, Logistic and Bertalanffy models showed the best corrected coefficient of determination (R2), over 0.98, and mean prediction errors (PE) were lower statistycally (P<0.05), in absolute value. The cubic regression provided also a high R-2, however, for male broilers the absolute value of PE was higher than the last three regressions (P<0.05). The linear followed by Richards and quadratic functions, provided poor estimates of observed weight in all studied periods. The Logistic and Gompertz functions provided adequate estimates of weigh and age at inflection point and daily gain rate. The Logistic function showed a 30% greater maturing rate than Gompertz, and the Bertalanffy over estimated the weight at maturity.