RUMEN DEGRADABILITY OF CRUDE PROTEIN OF DRIED GRASS AND LUCERNE FORAGE MEASURED BY IN-SACCO INCUBATION AND PREDICTED BY NEAR-INFRARED SPECTROSCOPY

Samples of grass (16) and lucerne (38) were harvested in advancing mat urity and growth stages within 1 year from April to October and were u sed to test the applicability of near infrared reflectance spectroscop y (NIRS) to estimate ruminal degradability of crude protein (CP). CP d egradability characteristics were measured with cows using the nylon b ag techniques. Changes in the nutrient contents were consistent with v egetation stage in the primary growth of grass and luceme and the firs t regrowth of luceme. Total degradability a+b was higher in lucerne fo rage than in grass. The means (+/- SE) for four growth cycles of lucer ne up to full bloom stage were 92.5 +/- 4.28, 93.9 +/- 2.05, 94.0 +/- 0.60 and 96.1 +/- 1.41, compared with 90.0 +/- 5.98 and 90.6 +/- 2.52 for grass up to anthesis in two cycles. Effective CP degradability (ED ) of both grass and lucerne decreased with increasing maturity. Up to late blooming stage ED values were between 70 and 82% for grass and 80 -90% for luceme. In the primary growth samples, linear relationships ( P < 0.001) were found between a + b and ED (Y) and the content of CP, crude fibre (CF) and the day of vegetation (X). Up to the end of bloom stage, for ED: R(2) ranged between 0.85 and 0.94 and SE between 0.78 and 2.9; for a + b the respective values were 0.89-0.97 and 0.59-3.2. NIRS was shown to be more universal than CP and CF content in predicti on of ED and a + b values of dried forage. Based on a common calibrati on at five wavelength terms for combined samples of grass and luceme ( n = 54), ED and a + b of all samples could be predicted with a mean st andard error of 2.59 and 3.27. NIR data accounted for 0.9 and 0.87 of variability in ED and a+b values in all samples of grass and luceme.