Effect of Composition of Feed Model Systems on Pellet Quality: A Mixture Experimental Approach. I

ABSTRACT Formulation is said to be the single most important variable affecting the physical quality of pelleted feeds. Previous investigations have usually focused on the influence of a single ingredient at one time and few have attempted to devise predictive models for pellet quality based on the composition of the rations. A study was conducted to investigate the functional properties of the compositional fractions of corn (starch, protein, fiber, fat) as related to the quality of pellets produced. Overall, a quadratic effect was observed for starch, with a positive response at lower‐ to midrange of inclusion, then assuming a downward curvature as the starch levels approached the higher end of the experimental region. Corn protein and especially fat appeared to have a negative influence on pellet quality. Empirical models were created that partially enabled the prediction of the response when ground yellow dent corn was substituted in the model systems. However, overestimation of the negative effects of fat in some of the models led to lower predicted values compared with actual observations.