The indigestible core of forage fiber: estimation from long-term in vitro kinetics and prediction from lignin content.

The indigestible core of forage fiber, quantified as indigestible neutral detergent fiber (iNDF, also denoted parametrically as Ω), critically influences ruminant intake, rumen fill, digestibility, and microbial substrate availability. This study estimated digestion kinetics and Ω for 17 forage species (nine grasses and eight legumes) using long-term in vitro incubations and developed a scaling function for Ω prediction from the Lignin(sa):aNDFom ratio. Multivariate analysis of chemical composition (DM, CP, CF, ash, aNDFom, Lignin(sa), and NFC) revealed distinct grass-legume profiles, with grasses showing greater aNDFom (592-763 g/kg DM) and lesser CP (76-166 g/kg DM) than legumes, and lignin ratios ranging 40-230 g/kg aNDFom. Incubations (up to 2,500 h, with medium and inoculum renewal every 96 h) produced biphasic degradation profiles that were fitted with nonlinear mixed models (EXP1, EXP2, GMM1, GMM2, and compartmental GηG1 and GηG1G1) in R (nlme). Information-theoretic criteria selected the heterogeneous two-pool compartmental model G6G1G1, which incorporated specimen-specific parameters, random block effects on the lag rate, power-of-the-mean variance, and a continuous autoregressive correlation structure. Fast pools ranged from 140-571 g/kg, slow pools from 59-379 g/kg, with corresponding degradation rates of 0.033-0.128 h⁻¹ and 0.0026-0.0102 h⁻¹. Parametric bootstrap (3,000 replicates) indicated that approximately 300 h of incubation is sufficient to capture 95% of potentially digestible fiber, despite uncertainty in the slow pools. The scaling law Ω = 2.22Lignin(sa)/aNDFom1.06 best predicted the indigestible residue, with diagnostics confirming model adequacy. Limitations include the modest number of forage specimens and residual uncertainty in the slow pools. This work integrates kinetic and chemical predictors to improve feed evaluation and ration formulation for ruminants.