Comparison of two in vitro procedures using rumen liquor‐pepsin or pepsin‐cellulase for prediction of forage digestibility

Abstract Seventy‐three samples of grass and legume herbage of known in vivo digestibility were used to compare an in vitro procedure comprising incubation with pepsin followed by cellulase (Jones and Hayward, 1975) with the Tilley and Terry (1963) procedure using rumen liquor followed by pepsin. The results obtained using the pepsin‐cellulase method proved to be highly repeatable. On the basis of the residual standard deviation of the regression of in vivo dry matter digestibility (DMD) on dry matter digested in vitro (g/100 g DM), the pepsin‐cellulase method was shown to be only slightly less accurate than the rumen liquor‐pepsin method for combinations of both primary growth and regrowth of grasses (r.s.d. ±1.80 and ±1.46 respectively). However, the pepsin‐cellulase method was noticeably less accurate than the rumen liquor‐pepsin method when used on legume herbages (r.s.d. ±3.17 and ±1.91) or on grass and legume herbages combined (r.s.d. ±3.80 and ±1.60). For the rumen liquor‐pepsin method one regression equation predicting DMD having the form, Y = 1.02 X− 0.41, was permissible. For the pepsin‐cellulase method separate regression equations were necessary for grasses (species and growth stages combined) having the form, Y = 0.56 X+ 34.7, and for each legume species. It was concluded that the pepsin cellulase method was of value for grass samples, and, within species, for legumes because of its accuracy and precision and its speed and convenience relative to the rumen liquor‐pepsin procedure; however, for legume species combined, and presumably for grass‐legume mixtures, it was not sufficiently accurate.