True Metabolizable Energy Values for Poultry of Commercial Diets Measured by Bioassay and Predicted from Chemical Data

Data describing 419 commercial poultry feeds and 41 mixtures of feedingstuffs were used to develop multiple linear regression equations for predicting the true metabolizable energy (TME) values of mixed feeds. Equations using the three-variable combinations "ether extract, crude fiber, ash" and "ether extract, crude protein, NFE" were equally useful as predictors of TME, explaining about 80% of the variation. The combination "ether extract, crude protein, carbohydrate", used in published equations for predicting apparent metabolizable energy, explained only about 60% of the variation. Equations with four or more variables were not significantly better than the three-variable equations. The commercial data were split into three groups. Equations developed on the separate groups were consistent. The equations are recommended for use with mixed feeds whose parameters fall within the ranges of the values observed in the experiment.