Anaerobic threshold as a basic criterion of training recommendation can be estimated by various parameters. The purpose of this study was to investigate the relationship and the reproducibility of ventilatory, lactate-derived and catecholamine thresholds of an incremental treadmill exercise. Therefore, 11 male subjects underwent two incremental treadmill tests within 7 days. The lactate threshold (LT) was determined at the lowest value of the lactate-equivalent (ratio lactate/performance). The individual anaerobic threshold (IAT) was calculated at LT+1.5 mmol/L lactate. The ventilatory thresholds, using mass-spectrometry, were defined by the V-slope method (AT) and at the deflection point of end-tidal CO2 (ET-CO2) concentration (RCP). The thresholds of epinephrine (TE) and norepinephrine (TNE) were calculated in the manner of LT. The running velocities were highly reproducible at LT (test-retest correlation coefficient r=0.90), IAT (r=0.97), AT (r=0.88) and RCP (r=0.95). By contrast TE (r=0.49) and TNE (r=0.46) showed a poor reproducibility. TE and TNE occurred 5-11% below LT and AT with a low correlation to LT and AT. LT was found 4% below AT, both were correlated with r=0.70 (p<0.01, test 1) and r=0.95 (p<0.01, test 2). IAT occurred 7-8% above RCP, in both tests a close correlation was found between IAT and RCP of r=0.97 (p<0.01). In summary, the ventilatory and lactate-derived thresholds show a high and similar reproducibility, but the catecholamine threshold does not. In the present exercise protocol, there are systematic differences between the lactate-derived and ventilatory thresholds, in spite of a close relationship, and these must be taken into account in recommendations derived for training.