Comparison between reconstituted sheep faeces and rumen fluid inocula and between in vitro and in sacco digestibility methods as predictors of intake and in vivo digestibility

This study (conducted in Debre Zeit, Ethiopia in 1993) examined (i) the effect of source of inoculum on in vitro dry matter (DM) digestibility (IVDMD) and gas production (GP) and (ii) the IVDMD, GP and in sacco degradability as predictors of in vivo DM digestibility (DMD) and intake. Six ruminally cannulated male sheep (used in the degradability studies and from which rumen fluid was harvested for the in vitro studies) and six intact sheep (from which faeces for reconstitution was obtained) were given teff straw ad libitum supplemented with 200 g/day of concentrate (1:1 mixture of noug (Guizotia abyssinica) cake and wheat middlings). In determining IVDMD on 26 feeds, pepsin-HCl digestion was replaced with neutral detergent extraction. Microbial GP was measured on these feeds incubated with rumen fluid or reconstituted faeces inocula at various time periods. The degradability of each feed was determined by the nylon bag technique in three sheep. Eighty-eight intact male Ethiopian Menz type sheep (mean liveweight 25.6 (S.D. = 1.98) kg) were used in a randomized complete block experiment to determine intake and digestibility. Gas production using faeces inoculum (GP-F) was strongly related to GP using rumen fluid inoculum (GP-R) particularly at 48 h (R(2) = 0 85; P < 0001) Of incubation. The IVDMD obtained with reconstituted sheep faeces inoculum (IVDMD-F) had a strong positive relationship (R(2) = 0.88; P < 0.0001) with IVDMD measured using rumen fluid (IVDMD-R). The IVDMD values obtained by centrifugation were positively related to those obtained by filtration for the rumen fluid (R(2) = 0.61) and reconstituted faeces (R(2) = 0.47) inocula. The relationship between degradability and in vivo DMD was generally poor for roughages but improved with the length of incubation for forage legumes, being strongest at 24 h (R(2) = 0.54). The in vivo DMD had the strongest relationship with GP-R at 24 h of incubation for roughages (R(2) = 0.64) and legumes (R(2) = 0.84). Dry matter degradability was poorly related to DM intake for roughages (R(2) < 0.10), while these were closely related for legumes particularly at 6 h of incubation (R(2) = 0.55). The relationship between GP-R and intake was strongest at 12 h of incubation for roughages (R(2) = 0.41) but was generally weak for legumes. The IVDMD accounted for a very low proportion of the variation in intake of roughages and legume forages. The gas production method was the best among the methods tested in predicting voluntary roughage intake, but not the intake of legumes. The nylon bag technique gave the best predictors (solubility, rate constant 'c' and Lag phase) of legume intake (R(2) = 0.98). The discussion suggests explanations for why the nylon bag technique predicted performance from forage legumes better than from roughages, why the technique was much better as a predictor of intake than of in vivo DMD, and why the gas production technique predicted in vivo DMD and not the intake of legumes. Reconstituted faeces inoculum may replace rumen fluid inoculum in in vitro procedures. The IVDMD method is a less accurate predictor of in vivo DMD than GP and rumen degradability constants. Both feed factors and practices inherent in the methodologies may modify the predictiveness of indirect digestibility methods.