Effects of varying bulk densities of steam-flaked corn (SFC) and level of inclusion of roughage in feedlot diets were evaluated in 3 experiments. In Exp. 1, a total of 128 beef steers were used in a 2 x 2 factorial arrangement to evaluate the effects of bulk density of SFC (335 or 386 g/L) and roughage concentration (6 or 10% ground alfalfa hay, DM basis) on performance and carcass characteristics. No interactions were observed between bulk density and roughage concentration for performance data. From d 0 to the end, cattle fed the 335 g/L SFC had greater overall G:F (P = 0.04) than those fed the 386 g/L SFC, with tendencies (P < 0.10) for improved G:F with the lighter flake weight evident at all 35-d intervals throughout the feeding period. Dry matter intake was less for cattle fed 6 vs. 10% roughage from d 0 to 35 (P = 0.03) and d 0 to 70 (P = 0.05), but not for the overall feeding period. Feeding 6 vs. 10% ground alfalfa as the roughage source tended (P = 0.09) to improve overall G:F. Treatment effects on carcass measurements were generally not significant (P > 0.20). In Exp. 2, the effects of bulk density of SFC (283, 335, or 386 g/L) and 6 or 10% ground alfalfa hay on IVDMD and in vitro pH were evaluated at 6, 12, 18, and 24 h of incubation. With a reduced-strength buffer in vitro fermentation system, pH increased (P < 0.01) with increasing bulk density at 6 and 12 h, and IVDMD decreased (P < 0.03) as bulk density increased. In contrast, in a normal-strength buffer system, there were no treatment differences (P > 0.23) for IVDMD. In Exp. 3, two diets that varied in bulk density of SFC and roughage concentration (335 g/L SFC with 6% alfalfa hay vs. 386 g/L SFC with 10% alfalfa hay) were compared for their effects on the pattern of feed intake and the acid-base balance in Holstein steers (12/treatment). No differences (P > 0.10) between treatments were noted for blood gases or urine pH; however, day effects (P < 0.02) were detected for blood pH, partial pressure of CO(2), and urine pH, which generally decreased (P < 0.05) with an increasing time on feed. The 2 treatments had little effect on the pattern of feed intake within the sampling days, with the exception that the 386 g/L SFC with 10% alfalfa hay diet increased (P < 0.05) the percentage of total DMI consumed at 1 and 6 h after feeding on d 14. Within the ranges of bulk density and roughage level studied, 335 g/L SFC with 6% alfalfa hay yielded the optimal animal performance, with limited effects on in vitro fermentation and the acid-base balance.