Breathing frequency (fb), tidal volume (VT), and respiratory timing during slow (SR, 8 W min-1) and fast (FR, 65 W min-1) ramp exercise to exhaustion on a cycle ergometer was examined in seven healthy male subjects. Expiratory ventilation (VE), pulmonary gas exchange (VO2 and VCO2) and end-tidal gas tensions (PET,O2 and PET,CO2) were determined using breath-by-breath techniques. Arterialized venous blood was sampled from a dorsal hand vein at 2 min intervals during SR and 30 s intervals during FR and analysed for arterial plasma PCO2 (PaCO2). PET,CO2 increased with increasing work rates (WRs) below the ventilatory threshold (VT); at WRs > or = 90% VO2,max, PET,CO2 was reduced (P < 0.05) below 0 W values in SR but not in FR.fb and VT were similar for SR and FR at all submaximal WRs, resulting in a similar VE. At exhaustion VE was similar but fb was higher (P < 0.05) and VT was lower (P < 0.05) in SR (fb, 51 +/- 10 breaths min-1; VT, 2590 +/- 590 ml) than in FR (fb, 42 +/- 8 breaths min-1; VT, 3050 +/- 470 ml). The time of expiration (TE) decreased with increasing WR, but there was no difference between SR and FR. The time of inspiration (TI) decreased at exercise intensities > or = VT; at exhaustion, TI was shorter (P < 0.05) during SR (0.512 +/- 0.097 s) than during FR (0.753 +/- 0.100 s). The TI to total breath duration (TI/TTot) and the inspiratory flow (VT/TI) were similar during SR and FR at all submaximal exercise intensities; at VO2,max, TI/TTot was lower (P < 0.05) and VT/TI was higher (P < 0.05) during SR (TI/TTot, 0.473 +/- 0.030; VT/TI, 5.092 +/- 0.377 l s-1) than during FR (TI/TTot, 0.567 +/- 0.050; VT/TI, 4.117 +/- 0.635 l s-1). These results suggest that during progressive exercise, breathing pattern and respiratory timing may be determined, at least at submaximal work rates, independently of alveolar and arterial PCO2.