Six males performed cycle ergometer exercise on two occasions in random order. Each exercise was preceded by a 2-h period in which matched capsules were administered orally, containing either starch (C) or NaHC03 (E) in a dose of a 0.2 g • kg-1 body wt; pre-exercise blood pH and [HC03 -] were 7.34±0.01 and 23.7±0.5 mM (mean + S.E.) for the C study, and 7.41±0.01 and 28.6 + 1.3 mM for the E study (p < 0.001 and p < 0.01, respectively). Exercise was continuous and maintained for 10 min at 40 of maximal oxygen uptake (40°c Vo z max)' followed by 15 min at 12W above the respiratory compensation threshold ([+ RCT]) which was determined by the increase of the ventilatory equivalent for carbon dioxide (VE V~oz 1), and for as long as possible at 95 % Voz max. Endurance time at 95 % Voz max was significantly longer in E than in C (2.98±0.64 min vs. 2.00 ± 0.44 min, p <0.05). The rate of increase in arterialized venous lactate (LA) was higher in E than in C from rest to exercise at [+ RCT], while there was no significant difference in the hydrogen ions ([H+]). Consequently, [H +] • LA 1 (nM . mM -1) was significantly lower in E than in C. The change of VE V~oz-' was shifted downward in E compared to C during exercise with the lowest value being observed at the same exercise stage. These results suggest that the respiratory responses to exercise are not affected by the higher level of [HC03 -] induced by NaHC03 ingestion, and appear to reflect the net change of plasma During intense exercise, lactic acid is formed and accumulates in the muscle. The decrease of muscle pH is one of the major factors limiting exercise performance (HERMANSEN and OsNES, 1972) by impairing enzyme activities (TRIVEDI and