Although many studies of the effect of dynamic exercise training on left ventricular (LV) mass have been reported, controversy continues to exist. Previous work has been criticized because of the techniques used for measuring LV mass, the variable level of training of the subjects recruited and the methods used to normalize the data. In an attempt to resolve this controversy, LV mass was determined using the very accurate and reproducible technique of magnetic resonance imaging (MRI). Highly trained competitive athletes including cross-country skiers, endurance cyclists and long distance runners (VO2max = 77 +/- 1, 72 +/- 2 and 75 +/- 2 ml (kg X min)-1, respectively) were examined. The data were normalized for body weight, body surface area and lean body mass. LV mass was significantly greater in skiers (239 +/- 9 g), runners (244 +/- 10 g) and cyclists (258 +/- 11 g) when compared with nonathletic control subjects (189 +/- 6 g) (p less than 0.001), which represents percent differences of 26, 29 and 37%, respectively. LV mass remained greater in the athletes, regardless of the method used to normalize the data. In addition, there was a good correlation between LV mass and VO2max (r = 0.80, p less than 0.001). It was concluded that LV mass is significantly greater in highly trained competitive endurance athletes and that normalizing LV mass with respect to body weight, body surface area or lean body mass does not alter this relation.