The relationship between dietary intake and urinary excretion of nitrate was investigated among 35 male and 24 female graduate students in Boston. The dietary assessment method consisted of a self-administered semi-quantitative food frequency questionnaire currently used for large-scale epidemiological studies. Calculated mean daily nitrate intake was 1.83 mmol for men and 2.96 mmol for women; broccoli and green leafy vegetables accounted for 60% of the total. Urinary measurements involved two overnight specimens with a mean collection time of approximately 13 hours. The ratio of intra-to-inter individual variance in urinary nitrate excretion (lambda) was 1.87. The simple correlation coefficient between intake and excretion of nitrate was found to be 0.20; after correction for the within-person variation by using lambda, the correlation coefficient was 0.28. Adjustment for gender, age and Quetelet's index in multiple regression analyses resulted in a partial correlation coefficient between nitrate intake and excretion of 0.37 (p = 0.005). Correction for within-person variation in urinary excretion increased this partial correlation coefficient between intake and excretion to 0.59 (95% CI = 0.03 to 0.87). These data suggest that a self-administered questionnaire may provide useful information on usual nitrate intake, and indicate the need to pursue this possibility further.