Projecten per jaar
For 10 batches of fresh grass, 11 of grass silage and 7 of grass hay, rumen degradation kinetics (soluble, potentially degradable and undegradable fraction and the degradation rate) of the organic matter (OM), crude protein (CP) and neutral detergent fibre (NDF), as well as the digestibility of rumen escape protein in the intestines (EPdi) were measured with 3 or 4 cows with cannulae in the rumen and duodenum. Data were used to calculate the fermentable OM content (FOM), the fraction of protein escaping the rumen (EP), the true protein digestible in the small intestine (DVE) and the degraded protein balance in the rumen (OEB) according to the Dutch protein evaluation system. The nutritive quality of fresh grass and silage was generally better than that of hay. On average, the FOM content of fresh grass, silage and hay amounted to 618, 571 and 552 g/kg DM, respectively. The EP of hay was almost double that of fresh grass and silage (0.535, 0.287 and 0.244 respectively). The EPdi did not differ greatly among grassland products, with 0.849 for fresh grass, 0.794 for silage and 0.816 for hay. The DVE content of fresh grass was similar to that of hay (86 and 85 g/kg DM) and higher than that of silage (68 g/kg DM). The OEB content was similar and positive for both fresh grass and silage (24 and 26 g/kg DM), whereas that of hay was negative (-30 g/kg DM). In order to predict in situ values, regression equations with growth days, chemical parameters, OM digestibility (OMd) and the washable CP fraction (W-CP) were developed for the three grassland products separately, as well as pooled. Most significant variables were OMd, W-CP, CP and ash and separate equations for the three grassland products were more accurate than a general equation. Use of the best product-specific equations resulted in a residual standard deviation (RSD) for the 28 studied samples of 8 g/kg DM for FOM, 0.016 for EP, 6 g/kg DM for DVE and 6 g/kg DM for OEB. For EPdi, a general equation for the pooled grassland products gave an RSD of 0.042. With near-infrared reflection spectroscopy (NIRS), the number of samples was too small to develop product-specific calibrations. A global calibration for the three grassland products gave cross-validation errors of 20 g/kg DM, 0.057, 0.049, 9 g/kg DM and 14 g/kg DM, for respectively FOM, EP, EPdi, DVE and OEB, clearly less accurate than with laboratory measurements. (C) 2004 Elsevier B.V. All rights reserved.