1. Four sheep, each fitted with cannulas in the rumen and proximal duodenum, were given two diets (1390 g dry matter (DM)/d) consisting of lucerne (Medicago sutiva) pellets (24.2 g nitrogen/kg DM) plus pelleted reed canary grass (Phaluris arundinacea; 14.1 g N/kg DM) or chopped hay (11.8 g N/kg DM) at intervals of 2 h. Flow of duodenal digesta measured by reference to the markers 51Cr-EDTA and 103Ru-phenanthroline indicated a net gain of 5.8–7.5 g non-ammonia-N (NAN) between mouth and duodenum.
2. The proportion of microbial N in duodenal digesta N was estimated using 15N and 35S incorporation into bacteria and digesta. Two methods of analysis for 35S content, the Bird & Fountain (1970; B&F method) and the Mathers & Miller (1980; M&M method), were used. (15NH4)2S04 and Na235S04 were infused into the rumen for 3.5 d before and 4.0 d during sampling. A bacterial fraction was prepared from the fluid phases of sampled duodenal digesta and rumen contents by differential centrifugation. In addition, samples of ground canary grass and of lucerne were incubated in nylon bags in the rumen for 3–48 h during the infusion.
3. Each of the 35S analytical methods yielded similar values of 35S content of isolated rumen or duodenal bacteria, but there was more (P < 0.05) incorporation of 15N into rumen than into duodenal bacteria. Relative to values obtained using the M&M method and 15N incorporation, the B&F method for S analysis yielded higher (P < 0.05) estimates of microbial content of duodenal digesta from sheep given chopped reed canary grass.
4. 35S activity associated with washed nylon-bag residues increased rapidly with time-period of incubation and was substantially greater (P < 0.05) when analysed by the B&F method compared with the M&M method. The 35S content (/g DM) of adherent bacteria removed from nylon-bag residues by homogenization in a second experiment vaned from 0.65 to 1.88 that of free-living bacteria isolated from rumen fluid by differential centrifugation.
5. The difference in 35S content in digesta and nylon-bag residues as measured using the two analytical methods was considered in relation to 35S-labelled extracellular material postulated to be produced by bacteria adherent to plant residues. Estimates of disappearance of dietary N from nylon bags after correction for microbial contamination indicated a disparity with estimates based on in vivo information.