Six ruminally and duodenally cannulated sheep were used in a partially replicated 4 x 4 Latin square to evaluate the effects of 4 diets on microbial synthesis, microbial populations, and ruminal digestion. The experimental diets had forage to concentrate ratios (F:C; DM basis) of 70:30 (HF) or 30:70 (HC) with alfalfa hay (A) or grass hay (G) as forage and were designated as HFA, HCA, HFG, and HCG. The concentrate was based on barley, gluten feed, wheat middlings, soybean meal, palmkern meal, wheat, corn, and mineral-vitamin premix in the proportions of 22, 20, 20, 13, 12, 5, 5, and 3%, respectively (as-is basis). Sheep were fed the diets at a daily rate of 56 g/kg of BW(0.75) to minimize feed selection. High-concentrate diets resulted in greater (P < 0.001) total tract apparent OM digestibility compared with HF diets, but no differences were detected in NDF digestibility. Ruminal digestibility of OM, NDF, and ADF was decreased by increasing the proportion of concentrate, but no differences between forages were detected. Compared with sheep fed HF diets, sheep receiving HC diets had less ruminal pH values and acetate proportions, but greater butyrate proportions. No differences among diets were detected in numbers of cellulolytic bacteria, but protozoa numbers were less (P = 0.004) and total bacteria numbers tended (P = 0.08) to be less for HC diets. Carboxymethylcellulase, xylanase, and amylase activities were greater for HC compared with HF diets, with A diets showing greater (P = 0.008) carboxymethylcellulase activities than G diets. Retained N ranged from 28.7 to 37.9% of N intake and was not affected by F:C (P = 0.62) or the type of forage (P = 0.31). Microbial N synthesis and its efficiency was greater (P < 0.001) for HC diets compared with HF diets. The results indicate that concentrates with low cereal content can be included in the diet of sheep up to 70% of the diet without detrimental effects on ruminal activity, microbial synthesis efficiency, and N losses.