Steers, fitted with simple rumen and duodenal cannulas, were given diets of approximately equal parts of flaked maize and hay (A) (containing 16 g nitrogen/kg dry matter (DM) or diets of flaked maize and straw supplemented with decorticated groundnut meal (B), fish meal (C), or heated (D) or unheated (E) soya-bean meal (containing 24 g N/kg DM). Chromic oxide was given as a marker with the feeds, and flows of different combined sugars at the duodenum estimated from the values for sugars:marker at this site. Contributions of bacterial constituents to these flows were estimated from amounts of RNA present. 2. Rumen bacteria from steers receiving diet A contained approximately 110 g alpha-dextran-glucose/kg DM and contributed about 60 g alpha-dextran-glucose/d at the duodenum; bacteria from steers receiving diets B, C, D or E contained 25-40 g alpha-dextran-glucose/kg DM and contributed about 20-30 g alpha-dextran-glucose/d at the duodenum. There were no significant differences between different N supplements. About half the alpha-dextran-glucose, varying amounts of the mannose and galactose, and nearly all the rhamnose and ribose in duodenal digesta were contributed by the bacteria. Almost all the arabinose, xylose and cellulose-glucose was of dietary origin. 3. For steers receiving diet A, mean coefficients of digestibility between mouth and duodenum, corrected where necessary for bacterial contribution, were 0-96, 0-73, 0-58, 0-22 and 0-53 for starch-glucose, galactose, arabinose, xylose and cellulose-glucose respectively. Corresponding mean values when diets B, C, D and E were given, which did not differ significantly amongst themselves, were 0-98, 0-79, 0-81, 0-59 and 0-58. Most values for galactose, arabinose and xylose were significantly higher than the diet A values.
