Eight healthy volunteers performed gradational tests to exhaustion on a mechanically braked cycle ergometer, with and without the addition of an inspiratory resistive load. Mean slopes for linear ventilatory responses during loaded and unloaded exercise [change in minute ventilation per change in CO2 output (delta VE/delta VCO2)] measured below the anaerobic threshold were 24.1 +/- 1.3 (SE) = l/l of CO2 and 26.2 +/- 1.0 l/l of CO2, respectively (P greater than 0.10). During loaded exercise, decrements in VE, tidal volume, respiratory frequency, arterial O2 saturation, and increases in end-tidal CO2 tension were observed only when work loads exceeded 65% of the unloaded maximum. There was a significant correlation between the resting ventilatory response to hypercapnia delta VE/delta PCO2 and the ventilatory response to VCO2 during exercise (delta VE/delta VCO2; r = 0.88; P less than 0.05). The maximal inspiratory pressure generated during loading correlated with CO2 sensitivity at rest (r = 0.91; P less than 0.05) and with exercise ventilation (delta VE/delta VCO2; r = 0.83; P less than 0.05). Although resistive loading did not alter O2 uptake (VO2) or heart rate (HR) as a function of work load, maximal VO2, HR, and exercise tolerance were decreased to 90% of control values. We conclude that a modest inspiratory resistive load reduces maximum exercise capacity and that CO2 responsiveness may play a role in the control of breathing during exercise when airway resistance is artificially increased.