The ruminal degradation of RNA in rye grass (Lolium perenne) was studied using the bag method. A non-lactating cow (BW 550 kg) fitted with a rumen cannula was used and fed twice daily at maintenance level with a chopped grass hay-based ration containing 30% ground barley. Rye grass, labelled during growth by fertilization with 15N2-urea (9.5 atom% 15N, 20 g N/m2), was cut at seven stages of growth and maturity and freeze-dried. RNA-N represented 6 to 17% of total N. Labelled grass samples (milled to 5.0 mm screen, 5.0+/-0.1 g DM) were incubated in polyester bags (100 x 200 mm, pore size 50 microm) in the rumen for periods of 1, 3, 6, 9, 12, 24, and 48 h. Data of N and RNA disappearances from the bags were fitted to an exponential equation to estimate parameters of degradation. The effective degradability of RNA in the rumen averaged 90+/-4%, for N it was 11% units lower (P < 0.001). Degradability of RNA was correlated to that of N (R2 = 0.92). Degradability of RNA (R2 = 0.96) and N (R2 = 0.93) decreased with increasing fibre content of grass. Increasing the fibre content by 1% diminished the degradability of RNA and N by 1.1% units and 2.4% units, respectively (P < 0.001). Assuming a microbial protein synthesis in the rumen of 150 g/kg DOM, a N: RNA ratio of 1:1.35 in rumen microbes and a rumen outflow rate of 0.06 h(-1), a model calculation indicates that about 9 to 19% of duodenal RNA are of dietary origin in animals fed grass. This should be taken into account for the calculation of microbial N on the basis of RNA as marker.