To evaluate group differences in coronary risk which could be attributed to the modality of habitual exercise, selected physiologic and lipid indices of coronary artery disease (CAD) were measured in 57 endurance trained (ET), strength trained (ST), or sedentary (SED) men (19 per group, aged 21 to 44 yr). Initial data reduction accomplished with principle component analysis identified three factors with eigenvalues greater than one. Orthogonal rotation of the preliminary solution demonstrated that low density lipoprotein cholesterol (LDL-C), percent body fat (%BF) and VO2max, and high density lipoprotein cholesterol (HDL-C) could be used to represent Factors 1, 2, and 3, respectively. The subsequent MANOVA using these variables proved significant. Post hoc analysis via simultaneous confidence intervals indicated that LDL-C group differences were not significant. Values for %BF and HCL-C in the ST group (14.0% and 1.17 mmol.l-1, respectively) were between but did not differ significantly from respective values in the ET (11.8% and 1.34 mmol.l-1) and SED (18.7% and 1.13 mmol.l-1) groups. However, %BF and HDL-C differences between the ET and SED groups were significant. The VO2max of the ET subjects (63.2 ml.kg-1.min-1) was significantly higher than that of either the ST or SED subjects (49.5 and 46.7 ml.kg-1.min-1, respectively). These results suggest that ET is the most effective modality of exercise for CAD risk reduction while benefits derived from ST are minimal.