Influence of fermentation methods on neutral detergent fiber degradation parameters.

The effect of 3 fermentation methods, in situ (IS) in 4 lactating cows (average pH of 5.8), in vitro (IVn) with media pH of 6.8, or in vitro (IVa) with media pH adjusted to 6.0 using citric acid, on fiber degradation parameters was studied using feeds ground to different particle sizes. Corn silage (CS), grass silage (GS), barley grain (B), sugar beet pulp (BP), and rapeseed cake (RC) were ground using a shear mill. Silages were ground through 8-, 4-, 2-, or 1-mm screens, B and BP through 4-, 2-, or 1-mm screens, and RC through 2- or 1-mm screens. The amylase-treated NDF (aNDF) content of samples ground using a 1-mm screen was 399, 431, 197, 480, and 251 g/kg of DM for CS, GS, B, BP, and RC, respectively, but increased with increasing screen size. Materials were incubated for 0, 6, 12, 24, 48, and 96 h IS, IVn, or IVa. Inoculum for IVn and IVa was prepared as a composite from the cows used for IS. The potentially degradable aNDF (D(0)), indegradable aNDF (I), lag time (L), and fractional rate of degradation of potential degradable aNDF (k(d)) were estimated using PROC NLIN in SAS. Except for RC, fermentation methods affected most degradation parameters, especially k(d) and L. The IVn, IVa, and IS methods resulted in k(d) values of 0.291, 0.105, 0.080 h(-1) and 0.262, 0.107, 0.103 h(-1) for BP and RC, respectively, demonstrating a decreasing rate of degradation for these feeds when fermented under suboptimal pH. In CS, GS, and B, no difference was found in k(d) between the IVn and IVa methods, which suggests that differences in pH did not alter k(d) in vitro. The k(d) values obtained for CS, GS, and B were 0.058, 0.109, 0.168, and 0.028, 0.054, and 0.069 h(-1) for the IVn and IS methods, respectively, indicating that the IS method using cows fed at production levels can underestimate the potential rate of NDF degradation. Using the IVa method, L was 12.1, 9.1, 7.8, and 2.5 h for CS, GS, BP, and RC, respectively, which was higher than L obtained from the IVn and IS methods for all feeds except B, where L in all methods were near the parameter boundary of zero hour in NLIN. It was concluded that fermentation methods were more important than grinding screen size on estimates for feed aNDF degradation and that the individual aNDF degradation parameters for the 5 feeds were affected differently by fermentation methods.