To elucidate patterns of dietary sodium ingestion in free-living subjects, we collected 26 consecutive 24-hour urine specimens in 18 subjects who had not received instructions to limit the sodium content of their diets, and who were not aware that sodium intake would be estimated from the collections. Nine subjects with plasma creatinine values less than 2 mg/dL had a mean 24-hour UNaV of 156 mEq/d, with an interindividual variability of SD +/- 38 mEq. Their intraindividual variability was 61 mEq. Nine subjects with creatinine values greater than 2 mg/dL had a mean 24-hour UNaV of 108 mEq/d (P less than 0.05). The intraindividual variability of these subjects was 39 mEq/d. Subjects with normal renal function ingested more sodium than subjects with renal insufficiency, although the variability in both groups was extensive. These data confirm and extend earlier observations and illustrate the difficulty in identifying biologic correlations in the presence of considerable intraindividual variability. They underscore the futility of estimating mean dietary sodium intake with a single or occasional 24-hour urine collection in both normal subjects and patients with renal insufficiency.