OBJECTIVES: To examine the association between physical activity and inflammatory markers, with consideration for body fatness and antioxidant use. DESIGN: Cross-sectional study, using baseline data from the Health, Aging and Body Composition Study. SETTING: Metropolitan areas surrounding Pittsburgh, Pennsylvania, and Memphis, Tennessee. PARTICIPANTS: Black and white, well-functioning men and women (N=3,075), aged 70 to 79. MEASUREMENTS: Interviewer-administered questionnaires of previous-week household, walking, exercise, and occupational/volunteer physical activities. Analysis of covariance was used to examine the association between activity level and serum C-reactive protein (CRP), interleukin-6 (IL-6), and plasma tumor necrosis factor alpha (TNFalpha) with covariate adjustment. Antioxidant supplement use (multivitamin, vitamins E or C, beta carotene) was evaluated as an effect modifier of the association. RESULTS: Higher levels of exercise were associated with lower levels of CRP (P180 min/wk). Adjustment for body fatness attenuated the associations somewhat. Use of antioxidant supplements modified the CRP (P-interaction=.01) and IL-6 (P-interaction=.08) associations such that concentrations were low in those taking supplements (e.g., CRP=1.79-1.84 across exercise levels) and higher in nonsupplement users who did no exercise (2.03) than in those who did the most (1.72). Among nonexercisers, higher levels of other physical activity were related to lower levels of CRP (P