Ten collegiate rowers performed discontinuous incremental exercise to their tolerable limit on two occasions: once on a rowing ergometer and once on a treadmill. Ventilation and pulmonary gas exchange were monitored continuously, and blood was sampled from a venous catheter located in the back of the hand or forearm for determination of blood lactate ([La]) and plasma epinephrine ([Epi]) and norepinephrine ([NE]) concentrations. Thresholds for lactate (LT), epinephrine (Epi-T), and norepinephrine (NE-T) were determined for each subject under each condition and defined as breakpoints when plotted as a function of O2 uptake (VO2). For running, LT (3.76 +/- 0.18 l/min) was lower (P < 0.05) than Epi-T (4.35 +/- 0.14 l/min) and NE-T (4.04 +/- 0.19 l/min). For rowing, LT (3.35 +/- 0.16 l/min) was lower (P < 0.05) than Epi-T (3.72 +/- 0.22 l/min) and NE-T (3.70 +/- 0.18 l/min) and was lower (P < 0.05) than LT for running. Within each mode of exercise, Epi-T and NE-T did not differ. Because LT occurred at a significantly lower VO2 than either Epi-T or NE-T, we conclude that catecholamine thresholds, per se, were not the cause of LT. However, for both modes of exercise LT occurred at a plasma [Epi] of approximately 200-250 pg/ml (rowing, 221 +/- 48 pg/ml; running, 245 +/- 45 pg/ml); these concentrations are consistent with the plasma [Epi] reported necessary for eliciting increments in blood [La] during Epi infusion at rest. Plasma [NE] at LT differed significantly between modes (rowing, 820 +/- 127 pg/ml; running, 1,712 +/- 217 pg/ml).(ABSTRACT TRUNCATED AT 250 WORDS)