1. Three sheep fitted with rumen and oesophageal cannulas were given hourly the following diets in successive experiments: 1000 g pelleted-grass cubes/d (diet A) and 700 g pelleted-grass cubes plus 300 g flaked barley/d (diet B).
2.During the final week of each 4-week dietary regimen, a balance trial was conducted and on separate days each sheep was continuously infused with [14C]urea and NaH14CO3 intravenously and NaH14CO3 intraruminaily. C transfer rates between blood urea, blood bicarbonate and rumen bicarbonate pools were calculated from the specific radioactivity of urea and bicarbonate sampled and isotope infusion rate during each experimental period. In the same period, an oral infusion of 51Cr-EDTA was maintained and salivary flow rate and composition determined from samples collected from the oesophageal fistula.
3. The inclusion of flaked barley in the pelleted-grass diet significantly (P < 0·01) increased the apparent digestibility of organic matter (0·069), apparently digestible organic matter intake and nitrogen balance, and increased the efficiency cf dietary N utilization from 0·059 (diet A) to 0·290 (diet B). Increased N balance was the result of a significant (P) < 0·01 reduction in urinary urea excretion.
4. The rumen fluid of sheep given diet A had higher pH and bicarbonate concentrations but lower butyric acid concentrations than that of sheep given diet B. There was no significant effect of diet on total volatile fatty-acid or ammonia concentrations in rumen fluid, or on osmolaiity and rumen fluid dilution rate. The irreversible loss of bicarbonate from rumen fluid was markedly increased when flaked barley was included in the diet, with most of the loss occurring directly from rumen fluid.
5. Sheep given diet A had higher salivary secretion rates (18·8 l/d) than those given diet B (12·7 l/d), and with the exception of urea, there was no effect of diet on the concentrations of total N, protein N, alpha;-amino-N, uric acid-N or bicarbonate in saliva. Urea concentrations in saliva were significantly correlated (r20·64) with blood urea concentrations, but not with salivary flow rate. Salivary secretions contributed 2·2 and 1·4 gN/d to the rumen of sheep given diets A and B respectively, with urea forming only 45–33% of the total N secreted.
6. When flaked barley was included in the pelleted-grass diet, there was a significant (P < 0·01) decrease in urea synthesis rate (diet A 20·0 g N/d, diet B 9·7 g N/d), a significant increase in amount (diet A 2·3 g N/d. diet B 3·0 g N/d) and proportion (diet A 0·024, diet B 0·57) of recycled urea degraded in the rumen. The permeability of the rumen wall to urea was also significantly increased in sheep given the flaked barley diet (diet A 1·35 g N/d, diet B 2·45 g N/d).
7. A model of urea metabolism in sheep given each diet is described, and the mechanisms by which flaked barley inclusion increased urea recycling to the rumen and the efficiency of dietary N utilization are discussed.