The percentage of feed protein escaping rumen degradation was measured by the in situ method (%EPsitu) for 29 compound feeds, untreated and formaldehyde-treated soyabean meal and 12 forages: 3 grass silages, 2 maize silages, fresh grass, grass hay, fodder beets, fresh potatoes, ensiled beet pulp, chopped ear-maize silage and brewers' grains. Loss of particles through bag pores was determined by the difference between the washable fraction (W) and the fraction soluble in borate-phosphate buffer at pH 6.7 (S). W - S was most pronounced for compound feeds (on average 14.4 percentage units), for brewers' grains and maize silages. A correction of %EPsitu, assuming that W - S degrades like the potentially degradable fraction, was not appropriate. Solubility in borate-phosphate buffer after 1 h, enzymic degradability by protease from Streptomyces griseus or ficin after 1, 6 and 24 h and near infrared reflectance spectroscopy (NIRS) (for compound feeds alone) were examined as a routine method to predict %EPsitu. With the buffer and S. griseus the effect of pH (6.7 vs. 8.0) and at pH 8.0 the effect of amount of substrate (500-mg sample vs. 20 mg N) were tested. With ficin, 500-mg samples were incubated at pH 6.7. Predictions were better when compound feeds and forages were considered separately. However, the best in vitro method was different for the 2 feed categories, being solubility in buffer for the compound feeds and enzymic degradation of a constant amount of protein with S. griseus at pH 8.0 for forages. NIRS showed potential to predict %EPsitu of compound feeds, but needs more reference samples. The Dutch feed tables appeared more accurate than the best in vitro method for compound feeds, but was too inaccurate for some forages like fodder beets, maize silage and ear-maize silage.
Prediction of the nutritive value of maize silage usingin vitroand near infrared reflectance spectroscopy (NIRS) techniques