The contribution of changes in body fatness and aerobic fitness to changes in blood lipids after aerobic exercise training was investigated. The sample included 295 men (77 black, 218 white) and 355 women (131 black, 224 white), aged 17 to 65 years, from the HERITAGE Family Study. Participants underwent measurements at baseline and after 20 weeks of supervised exercise training on a cycle ergometer. Body fat mass (FM, in kilograms) was determined by underwater weighing, and aerobic fitness (maximal oxygen uptake, VO(2max), in milliliters per minute) was assessed by cycle ergometry. Blood lipid measurements included fasting plasma levels of high-density lipoprotein cholesterol (HDL-C), HDL(2)-C, HDL(3)-C, low-density lipoprotein cholesterol (LDL-C), total cholesterol (CHOL), CHOL/HDL, and triglycerides (TG). A composite lipid change index (LCI) was derived by subjecting the Delta scores for the individual blood lipids to principal components analysis. The exercise training was accompanied by a mean increase of 17.5% in VO(2max) and a mean decrease of 3.3% in FM. Partial correlations, controlled for age, between absolute changes in VO(2max) and changes in the blood lipids were consistently low and nonsignificant. On the other hand, absolute changes in FM were significantly (P <.05) associated with changes in HDL-C (r = -.23), HDL(2)-C (r = -.17), and CHOL/HDL (r =.24) and the LCI (r = -.27) in men and with changes in LDL-C (r =.22), CHOL (r =.19), and CHOL/HDL (r =.15) and the LCI (r = -.19) in women. Forward stepwise regression confirmed that the change in FM was a better predictor of changes in blood lipids than the change in VO(2max), entering as a predictor in 4 of 8 regressions in both men and women. Change in VO(2max) did not enter as a significant predictor in any regression. Further, there were no differences in LCI between the upper and lower quartiles of VO(2max) change. On the other hand, there were significant differences between the low and high quartiles of FM change. No race effects were observed in any of the relationships, except that race was a significant predictor of changes in TG in both men and women. In conclusion, changes in blood lipids associated with aerobic exercise training do not appear to be related to changes in aerobic fitness per se; rather, they are weakly to moderately associated with changes in body fatness.