We studied separately the effects of weight-loss by dieting or by running on apolipoprotein (apo) A-I, apo A-II, and high-density lipoprotein (HDL) subfractions in sedentary, moderately overweight men assigned at random into three groups: exercise without calorie restriction, calorie restriction without exercise, and control. The absorbance of protein-stained polyacrylamide gradient gels was used as an index of mass concentrations for five HDL subclasses that have been identified by their particle sizes: HDL 3c (7.2 to 7.8 nm), HDL 3b (7.8 to 8.2 nm), HDL 3a (8.2 to 8.8 nm), HDL 2a (8.8 to 9.7 nm), and HDL 2b (9.7 to 12.8 nm). During the l-year trial, the exercisers ran (mean ± SD) 15.8 ± 8.1 km/wk, and the dieters reported eating 340±71 fewer calories per day than at baseline. Total body weight and fat weight were both reduced significantly more in dieters (−7.2 ± 4.1 and −6.2 ± 4.1 kg, respectively) and in exercisers (−4.0 ± 3.9 and −4.6 ± 3.5 kg) than in controls (0.6 ± 3.7 and −0.7 ± 2.7 kg). As compared with mean changes in controls, exercisers and dieters each decreased HDL 3b and increased HDL 2b . Exercisers also significantly increased plasma apo A-I concentrations. Analysis of covariance was used to statistically adjust the mean lipoprotein changes for the effects of weight-loss. The adjustment eliminated the significant reductions in HDL 3b and the significant increases in HDL 2b in exercisers and dieters, and it eliminated the significant increase in apo A-I in exercisers. When adjusted, the dieters' mean changes in HDL 2b had significantly decreased relative to those of both exercisers and controls. These results suggest that the changes in HDL 3b , HDL 2b , and apo A-I during 12 months of exercise could be due to metabolic changes associated with exerciseinduced weight-loss. Reduced calorie flux or other factors appear to attenuate the increase in HDL 2b and apo A-I during diet-induced weight-loss.Running increases the plasma mass and cholesterol concentrations of high-density lipoprotein (HDL) 2 in men [1][2][3]. The HDL 3 response to running is less clear, with some studies showing an increase and other studies showing either a decrease or no change [1,2,4]. HDLs may be further divided into two HDL 2 and three HDL 3 subclasses by nondenaturing polyacrylamide gradient gel electrophoresis [5]. They are defined by their estimated particle diameters: HDL 3c (7.2 to 7.8 nm), HDL 3b (7.8 to 8.2 nm), HDL 3a (8.2 to 8.8 nm), HDL 2a (8.8 to 9.7 nm), and HDL 2b (9.7 to 12.9 nm) [5]. These HDL subclasses show relationships to other lipoprotein variables that suggest that HDL 2a and HDL 2b are inversely related, and HDL 3b is positively related to coronary heart disease risk [6]. To our knowledge, the effects of exercise or diet on HDL subclass measurements by gradient gel electrophoresis have not been previously reported. This report describes the effects of diet-induced weight-loss and exercise-induced weight-loss on plasma apolipoprotein (apo) A-I and apo A-II concentrations, and on protein-staine...