Few studies provide correlations between different indicators of the dietary intakes of older (>or= 65 y) black and white adults. This study compared the usual intakes of vitamin E, vitamin C, and selected carotenoids estimated by a modified Harvard food frequency questionnaire (FFQ) to those estimated by multiple 24-h recalls, and to blood concentrations of components in a randomly selected sample of participants in the Chicago Health and Aging Project (CHAP). Subjects (n = 59) were interviewed to complete multiple 24-h recalls over a year's time, then completed an FFQ and subsequently provided a fasting blood specimen within 2 mo. Dietary estimates were energy-adjusted separately for men and women. Significant (P < 0.05) correlations between total (diet and supplement) FFQ estimates and serum measures of vitamin E and vitamin C were as follows: 0.49 and 0.39 for blacks, and 0.42 and 0.29 for whites, respectively. The highest correlations between serum and FFQ indicators were for dietary beta-cryptoxanthin (0.46), total alpha-tocopherol (0.46) and total beta-carotene (0.44) among whites; among blacks, the highest correlations were for dietary alpha-carotene (0.81), total alpha-tocopherol (0.53) and total beta-cryptoxanthin (0.50); all were significant (P < 0.05). Further adjustment for age, gender, BMI, and educational level minimally altered these coefficients. These findings indicate that the modified Harvard FFQ provides reasonable estimates of serum levels of vitamin E, vitamin C and beta-cryptoxanthin among CHAP participants.