To determine the influence of training and detraining on left ventricular dimensions, echocardiographic estimates of left ventricular indexes were undertaken in two groups of young healthy subjects. The training group consisted of eight competitive swimmers who were studied serially for 9 weeks. Left ventricular end-diastolic dimension in this group increased from the pretraining value of 48.7 ± 1.7 (mean ± standard error) to 53 ± 0.2 mm by the 1st week and to 52 ± 1.7 mm by the 9th week of training (P <0.005). Left ventricular posterior wall thickness increased gradually from 9.4 ± 0.4 to 10.1 ± 0.4 mm by the end of the training period (P < 0.005). The estimated derived indexes (left ventricular end-diastolic volume index, mass and stroke volume) also increased significantly. However, changes in ejection fraction were insignificant. Maximal oxygen uptake (VO2 max) increased from 52 ± 3 to 60 ± 3 ml/kg per min (P < 0.005). The detraining group comprised six competitive runners who stopped training for 3 weeks. Left ventricular end-diastolic dimension in this group decreased progressively from the control value of 51.0 ± 3 to 46.3 ± 2.3 mm (P <0.005) and posterior wall thickness from 10.7 ± 0.3 to 8.0 ± 0.7 mm (P <0.005) by the end of the detraining period. Likewise, the estimated derived indexes were significantly lower in response to detraining. Detraining did not influence ejection fraction. VO2 max decreased from 62 ± 2 to 57 ± 2 ml/kg per min (P <0.005). The results indicate that (1) exercise training-induced adaptive changes in left ventricular dimensions occur rapidly and mimic the pattern of chronic volume overload; and (2) modest degrees of exercise-induced left ventricular enlargement are reversible after cessation of training.