The main purpose of this study was to analyze the effects of exercise mode, training status and specificity on the oxygen uptake (V(.)O(2)) kinetics during maximal exercise performed in treadmill running and cycle ergometry. Seven runners (R), nine cyclists (C), nine triathletes (T) and eleven untrained subjects (U), performed the following tests on different days on a motorized treadmill and on a cycle ergometer: (1) incremental tests in order to determine the maximal oxygen uptake (V(.)O(2max)) and the intensity associated with the achievement of V(.)O(2max) ( IV(.)O(2max)); and (2) constant work-rate running and cycling exercises to exhaustion at IV(.)O(2max) to determine the "effective" time constant of the V(.)O(2) response (tauV (.)O(2)). Values for V(.)O(2max) obtained on the treadmill and cycle ergometer [R=68.8 (6.3) and 62.0 (5.0); C=60.5 (8.0) and 67.6 (7.6); T=64.5 (4.8) and 61.0 (4.1); U=43.5 (7.0) and 36.7 (5.6); respectively] were higher for the group with specific training in the modality. The U group showed the lowest values for V(.)O(2max), regardless of exercise mode. Differences in tau V(.)O(2) (seconds) were found only for the U group in relation to the trained groups [R=31.6 (10.5) and 40.9 (13.6); C=28.5 (5.8) and 32.7 (5.7); T=32.5 (5.6) and 40.7 (7.5); U=52.7 (8.5) and 62.2 (15.3); for the treadmill and cycle ergometer, respectively]; no effects of exercise mode were found in any of the groups. It is concluded that tau V(.)O(2) during the exercise performed at IV(.)O(2max) is dependent on the training status, but not dependent on the exercise mode and specificity of training. Moreover, the transfer of the training effects on tau V(.)O(2) between both exercise modes may be higher compared with V(.)O(2max).