A 4 × 4 Latin square design experiment with 3-week experimental periods was conducted with four wether sheep, each fitted with a permanent rumen cannula, to evaluate the effects of dietary protein sources and levels on fermentation and microbial activity in the rumen. Four complete diets were offered each containing (g/kg dry matter (DM)) molasses 248, grass silage 200 and barley straw 260. The control diet (C) also contained barley and soya-bean meal and the other three diets were supplemented with urea (CU), soya-bean meal (CS), and soya-bean meal and fish meal (CSF), respectively. This gave foods of similar concentrations of metabolizable energy (ME) and estimated fermentable ME (10·6 and 9·8 MJ/kg DM, respectively), but different levels (g/kg DM) of estimated effective rumen degraded dietary protein (ERDP) and digestible undegraded protein (DUP) (ERDP/DUP, 84/17, 109/17, 116/38 and 119/54 for diets C, CU, CS and CSF, respectively).
No clinical symptoms of ill health in the animals due to the feeding of molasses were observed during the experiment. The average pH values of rumen liquors obtained at various sampling times post feeding for diets C, CU, CS and CSF were 6·40, 6·49, 6·62 and 6·47 (s.e.d. 0·06 P < 0·05) respectively and average ammonia-nitrogen concentrations were 63, 81, 90 and 113 mg/l (s.e.d. 14·9, P < 0·02) respectively. The average concentrations of total volatile fatty acids in the rumen liquor were similar across the four treatments. The molar proportions of propionate and butyrate were higher for the diet C than for the other three diets (P < 0·05), while acetate was lower (P < 0·05). Supplementing with true protein (P < 0·05), but not with urea (P > 0·05), increased the molar proportions of isobutyrate and isovalerate. Whole tract apparent digestibilities of DM and organic matter did not differ significantly across the four treatments, but neutral-detergent fibre apparent digestibility (0·677, 0·672, 0·716 and 0·728 (s.e.d. 0·017) g/kg DM for diets C, CU, CS and CSF respectively) and the proportions of hay DM that disappeared in the rumen during 24 h incubation (0·223, 0·238, 0·284 and 0·271 (s.e.d. 0·019) g/kg DM) were significantly lower for diets C and CU than CS and CSF (P < 0·05). Urinary excretion of purine derivative nitrogen was similar across the four treatments. The results obtained from the present study indicate that there were no significant differences in the microbial crude protein synthesis in the rumen when a diet containing molasses was supplemented with urea or true protein. However, the supplementation of this control diet with true protein, but not with urea, did stimulate the degradation of hay DM in the rumen and the digestion of dietary fibre in the whole tract.