Four hundred thirty-two crossbred yearling steers (339 kg +/- 4.8) were used to investigate the effects of water quality and dietary potassium concentration and source on feedlot performance and carcass merit. The study was conducted using a 2 x 3 factorial treatment arrangement. Factors evaluated included 2 water sources: 1) a blend of reverse osmosis and well water (RO; 608 +/- 164 mg/L of SO(4)) and 2) 100% well water with SO(4) concentration of 1,933 +/- 53 mg/L and 3 dietary K treatments. Potassium treatments included 0.75% K with supplemental K from potassium chloride (0.75% K-KCl), 0.75% K with supplemental K from potassium carbonate (0.75% K-K(2)CO(3)), and 1.0% K with supplemental K from potassium carbonate (1.0% K-K(2)CO(3)). Interactions between water quality and dietary treatments were not significant. Dry matter intake tended (P = 0.10) to be greater for steers consuming RO water compared with well water and was not affected by dietary treatment. Feed efficiency (P = 0.04) and NE(g) recovery (P = 0.04) were greater for 1.0% K-K(2)CO(3) compared with 0.75% K-KCl but were not affected by water quality. Final BW was heavier (P < 0.001) and ADG was greater (P = 0.04) for RO water compared with well water but were not affected by diet. Carcasses from steers that consumed RO water tended (P = 0.08) to be heavier than carcasses from steers consuming well water. Dietary treatment did not affect HCW (P = 0.52). Yield grade calculated from carcass measurements was not affected by dietary treatment or water quality. Carcasses from steers consuming well water had greater (P = 0.04) marbling scores than RO water. These data demonstrate that steers consuming RO water achieved improved feedlot performance. Steers fed 1.0% K-K(2)CO(3) were more efficient and demonstrated improved energy recovery compared with steers fed 0.75% K-KCl. Improved efficiency and energy recovery may be related to a reduction (P = 0.06) in the liver abscess rate for steers consuming 1.0% K-K(2)CO(3). Dietary cation-anion balance was positively related to ADG (P < 0.01) and NE(g) (P = 0.03) recovery but negatively related to marbling score (P = 0.04).