We previously showed that African Americans excreted less urinary potassium than whites, even while consuming similar diets in the Dietary Approaches to Stop Hypertension (DASH) trial. We hypothesized that a low-sodium diet may eliminate these differences. Data from the DASH-Sodium randomized controlled feeding trial were analyzed. 412 adults with prehypertension or stage 1 hypertension. Random assignment to either a typical American "control" diet (1.7 g [43 mEq] potassium/2,100 kcal/d) or the DASH diet (4.1 g [105 mEq] potassium/2,100 kcal/d). Within each diet, participants received 3 levels of sodium intake in random order for 30 days. 24-hour urine samples were analyzed at the end of each period. The primary outcome was urinary potassium excretion. On the DASH diet, African Americans consistently excreted significantly less urinary potassium (mean 24-hour urinary potassium excretion, 2,594 ± 961 mg [66 ± 25 mEq]) than whites (3,412 ± 1,016 mg [87 ± 26 mEq]) at the highest sodium level; adjusted (P < 0.001); this difference was not altered by sodium level (P = 0.6 comparing white to African American difference in urinary potassium excretion on high- vs low-sodium diet). In contrast, there was a smaller but significant white-African American difference in mean daily urinary potassium excretion in participants fed the control/high-sodium diet that was not present in the control/low-sodium diet (adjusted differences of 281 mg [7 mEq]/d vs 20 mg [0.5 mEq]/d, respectively; P = 0.007). Significant interactions were found between race and diet (P < 0.001) and between race and sodium (P = 0.02). Single rather than multiple urine collections were available at each time. Lack of stool potassium and sweat potassium values. Racial differences in urinary potassium excretion depend on sodium intake and diet. Our results may help explain the previously documented large variability in urinary potassium excretion.