Comparison of Compartmental and Adams-Bell Models of Poultry Egg Production

Abstract

Weekly egg production data from six first cycle and 13 molted commercial layer flocks were used to compare three nonlinear egg production models: the compartmental or McMillan model, the Adams-Bell model, and a compartmental-type model based on a logistic growth curve. Models were fitted using the Marquardt method of the NLIN procedure [SAS Institute, Inc. (1985)] with results analyzed using a random block analysis of variance. Mean coefficient of determination (R2) values for the Adams-Bell (.9938) and the logistic (.9930) models were significantly higher than for the compartmental model (.9523) for hen-housed data from first cycle flocks, with no differences in molted flocks. The R2 values based on the 1st 24 wk of production followed the same pattern. Predictions of total production based on 24 wk of data were significantly more accurate for the Adams-Bell and logistic models than for the compartmental model in terms of error or percentage error in first cycle flocks, with no difference in molted flocks. Eliminating early data points where egg production was less than 20% significantly reduced the R2 value of the Adams-Bell model and significantly improved the R2 value of the compartmental model. The Adams-Bell model may be useful in decision making concerning replacement or molting of layer flocks.