The Prediction of the Nutritive Value of Animal Feeds from Chemical Analyses

Abstract

Summary A protein index for animal feeds is suggested as a basis for the criteria relating nutritive value to chemical composition. The feeds were classified into four main groups: legumes, non-legumes, straws and concentrates. The regression equations for predicting the nutritive value gave a reasonable level of accuracy (0·18–14·74% error). Higher levels of accuracy were attained employing regression equations formulated for more homogenous feed subgroups classified according to their protein and fibre contents. The percentage error resulting from the use of the specific equations was found to range between zero and 9·88.