1. In a 4 × 4 Latin square experiment four cattle were given in two meals per d diets consisting of (g/kg dry matter (DM)) 500 barley, 400 grass silage and 100 soya-bean meal. The diets were given at either 1.15 (L) or 2.3 times (H) maintenance energy requirements and the soya-bean meal was either untreated (U) or formaldehydetreated (T).
2. A 24 h collection of duodenal digesta and a 7 d collection of faeces were made using chromium sesquioxide for flow estimation and 35S as a marker of microbial nitrogen entering the small intestine. Samples of rumen fluid were also taken for estimation of lumen pH, ammonia and volatile fatty acid concentrations.
3. Spot samples of duodenal digesta were obtained after administration of Cr2O3,-mordanted silage-fibre and soya-bean meal, to determine the rates of outflow of these markers from the rumen. Similar samples were also obtained after cessation of a continuous intraruminal infusion of ruthenium phenanthroline, 35S and CoEDTA.
4. Incubations of each feedingstuff in porous synthetic fibre (psf) bags were carried out in the rumen and the rates of N disappearance from the bags determined.
5. Increasing DM intake significantly ( P < 0.001) increased the quantities of organic matter (OM), total N and amino acid-N entering the small intestine and amounts subsequently voided in the faeces. Apparent digestibilities of OM and N were unaffected by DM intake; the proportions of total digestible OM digested in the rumen were significantly lower (P < 0.01) at the higher level of DM intake.
6. Formaldehyde treatment of the soya-bean meal increased the quantities of N entering the small intestine; these increases were not significant.
7. Increased DM intake increased the quantities of both microbial N (P < 0.001) and undegraded feed N (P < 0.01) entering the small intestine; HCHO-treatment also significantly (P < 0.05) increased the quantities of undegraded feed N entering the small intestine. The efficiency of microbial N synthesis within the lumen was not significantly affected by dietary treatments whereas apparent feed N degradability was reduced significantly ( P < 0.05) both by increasing DM intake and by HCHO-treatment of the soya-bean meal.
8. Rates of disappearance of N from psf bags in the rumen were different for different feedingstuffs. However, for a given feedingstuff, the rate of N disappearance was not affected by the diets fed.
9. The rates of decline in marker concentrations measured in duodenal digesta were significantly increased as DM intake increased with the exception of Cr2O2-soya-bean meal. The markers could be ranked (P < 0.05) in the following order of increasing outflow rate: ruthenium phenanthroline, 35S-labelled amino acids and Cr2O2-silage fibre < Cr2O3-soya-bean meal < CoEDTA.
10. Estimates of the degradabilities of feedingstuffs were calculated from N disappearance rates from psf bags and either experimentally determined outflow rates or those proposed by the Agricultural Research Council (1984). Such estimates for the degradability of the whole diet were then compared with those determined in vivo using 35S as a marker.