The aim of this study was to examine the contribution of different anthropometric and metabolic indices to a rowing ergometer performance as determined by a maximal 2500m rowing ergometer (Concept II, USA) test. Twenty-one male experienced rowers of different boat categories participated in this study. The participants were also subjected to an incremental exercise test on a rowing ergometer to determine maximal oxygen consumption (VO2max), the corresponding maximal aerobic power (P-max), lactate threshold (LT) and rowing economy as the lactate (LA) concentration at the power of 350W (LA(350W)). Mean 2500m rowing ergometer performance time was 517.2+/-25.4s (8 min 37.2s) and the LA recovery parameters measured after the test were not significantly related to the rowing ergometer performance time. The LBM (kg) from the anthropometric and the Pm, (W) from the metabolic variables showed the highest correlation with the rowing ergometer performance time (r=-0.88). In addition, the shared variance was also greater than 50% (i.e. r(2)>0.50) between rowing performance time and body mass (kg), muscle mass (kg), skeletal mass (kg), CSA thigh (cm(2)), VO2max (l.min(-1)) and LA(350W) (mmol.l(-1)). A stepwise multiple regression analysis demonstrated that P-max was the best single predictor of 2500m rowing ergometer time, as a model incorporating P-max explained 77.3% of the total variance in rowing performance. The inclusion of muscle mass improved the prediction of 2500m rowing ergometer time, as the model with these two predictors explained 85.8% of the variance in rowing performance. The results of this study suggest that rowing performance is determined by the P-max and LBM indices.