The prediction of dry matter intake of temperate and tropical roughages from in vitro digestibility/gas-production data, and the dry matter intake and in vitro digestibility of African roughages in relation to ruminant liveweight gain

The dry matter intake (DMI) of 54 roughages has been predicted by various parameters obtained from the incubation in an in vitro gas-production system. Rate and extent of gas production from whole roughage and isolated NDF were described by two exponential models y = B(1 − e−c(t − lag)) and y = B(1 − e−ct) for t > lag, where y = gas volume at a given time t, B = potential cumulative gas production, c = rate of gas production and lag = initial lag for the onset of fermentation. Approximately 69% (P < 0.0001) of the variation in DMI of the 54 roughages was accounted for by the combination of B (59.6%), lag (5.6%) and c (3.4%). In the model where t > lag, B accounted for 53.6% of the variation in DMI and c for 16.1%. Approx. 81% (P < 0.0001) of the variation in DMI was explained by the kinetic parameters based on NDF incubations (BNDF 77.5%, CNDF 3.7%, lagNDF not significant). In the model where t > lag, BNDF accounted for 77.5% and cNDF for 5% of the variation in DMI. Combinations of in vitro true degradabilities at 36 or 48 h with gas volumes measured in the same incubation either at 4, 6 or 8 h accounted for approximately 82.0% of the variation in DMI and the inclusion of the partitioning factor (PF) improved the correlation (r2) to 0.87. The PF reflects the partitioning of truly degraded substrate between short chain fatty acids and microbial cells, and roughages having proportionally higher microbial yield had higher DMI. These parameters together also predicted liveweight gain (LWG) of sheep fed six roughages from Ethiopia (r2 = 0.98–0.99). The importance of DMI information for LWG is also discussed.