Twenty-one mildly obese women were fed a 1268-kcal lactovegetarian diet for 5 wk. Subjects were randomly divided into either an exercise (Ex) or a nonexercise (Nex) group. The Ex group walked at 60% heart-rate reserve (134 +/- 2 bpm; mean +/- SEM) during 45-min sessions, five times per week. Although exercise improved estimated maximum oxygen consumption (VO2max; 20.9 +/- 3.2% vs 2.1 +/- 3.4% in Ex vs Nex, respectively), changes in total body, lean body, and fat weight did not differ significantly between groups. Total body weight decreased 5.5 +/- 0.6 and 5.6 +/- 0.2 kg in Ex and Nex, respectively. Statistical analysis revealed significant differences in the pattern of change between groups for serum high-density-lipoprotein cholesterol (HDL-C; F[2,38] = 5.93, P = 0.006) but not for total cholesterol (TC), low-density-lipoprotein cholesterol, triglycerides, or glucose. Change in diet quality and body weight were found to account fully for the 12.7 +/- 1.9% decrease in TC. When change in dietary quality and body weight are equated in a group of mildly obese women, the effect of moderate exercise training on indices of serum lipid and lipoprotein is limited to HDL-C.