Prediction of laboratory and in situ protein fractions in legume and grass silages using near-infrared reflectance spectroscopy

Legume and grass silage samples (n = 121) were collected from commercial forage testing laboratories (trial 1). Samples were dried at 55 degrees C for 48 h, ground, scanned on a near-infrared reflectance spectrophotometer, and analyzed for crude protein (CP), soluble CP, acid detergent fiber (ADF) CP, and neutral detergent fiber (NDF) CP by wet chemistry methods. Sixty samples were selected for calibration development, and the remaining samples were used for equation validation. Near-infrared reflectance spectroscopy accurately predicted the CP content of the silages (R2 = 0.96), but prediction of soluble CP, ADF CP, and NDF CP was markedly less accurate. The coefficients of determination and standard errors of calibration for CP, ADF CP, NDF CP (percentage of DM), and soluble CP (percentage of CP) were as follows (0.96 and 0.80, 0.77 and 0.24, 0.72 and 0.71, and 0.82 and 4.40). In a second study, legume and grass silage samples (n = 32) were dried at 55 degrees C and ground (2 mm). Duplicate dacron bags containing 5 g of silage were incubated in the ventral rumen of three ruminally cannulated cows for 0, 3, 6, 12, 24, 48, and 72 h. In situ protein fractions, including rapidly degraded protein, slowly degraded protein, undegradable protein, degradation rate, and rumen-undegradable protein, were determined. Original samples were reground (1 mm) and scanned. Previously defined near-infrared spectroscopy calibration procedures were conducted. Coefficients of determination for in situ CP fractions were R2 > 0.92 with the exception of degradation rate (R2 = 0.87). Data suggest that in situ protein fractions are better predicted by near-infrared reflectance spectroscopy than by laboratory protein fractions.