Abstract
In vitro rumen liquor-pepsin, pepsin-cellulase and fibre analyses were compared as methods to predict the digestibility in vivo of each of four cuts of Setaria sphacelata, Chloris gayana and Pennisetum purpureum. Dry matter digestibility (DMD) determined by the rumen liquor-pepsin method was closely related to in vivo DMD (residual standard deviation (rsd) 1·65). The dry matter (DM) solubilized by cellulase enzymes from two sources, viz. Basidiomycetes and Trichoderma viride, predicted in vivo DMD with less accuracy than the rumen liquor-pepsin procedure (rsd 3·39 and 2·86 for Basidiomycetes and T. viride respectively). On average, T. viride enzyme solubilized 33 per cent more grass DM than did the enzyme derived from Basidiomycetes. The ashfree contents of cell wall (CW), acid detergent fibre (ADF) and lignin were relatively poor predictors of digestibility determined in vivo (rsd 3·97, 3·09 and 3·08 for CW, ADF and lignin respectively). The digestibility of cell contents was constant at 88 per cent but the proportion of lignin in the ADF accounted for only 38 per cent of the variance of the in vivo digestibility of the CW. Nevertheless, the fit of the latter tended to be improved by considering each species separately.
