The effects of supplemental oxygen on exercise tolerance and exercise-induced dyspnea in patients with chronic obstructive pulmonary disease (COPD) and an arterial oxygen tension (PaO2) over 60 Torr were studied. Twenty two men who met the criteria stated above underwent three 6-minute walk tests while breathing room air, compressed air, or supplemental oxygen, in a double-blind randomized crossover fashion. The distance walked in 6 minutes was recorded, and the degree of dyspnea during exercise was quantified on a visual analog scale with a questionnaire that had eight descriptors of dyspnea sensations. The distances walked were: 535 +/- 27 (SE) for room air, 545 +/- 27 for compressed air, and 563 +/- 23 m for O2. These data indicate a significant placebo effect and an effect of supplemental oxygen. The data regarding dyspnea also indicated both a significant placebo effect and an effect of supplemental oxygen, but the descriptors for the placebo effect differed from those for the effect of supplemental oxygen. Patients in whom arterial desaturation (below 88%) developed during exercise with compressed air benefited from supplemental oxygen: they walked significantly farther and their dyspnea scores for two of the eight descriptors were significantly lower. However, patients in whom arterial desaturation (below 88%) did not develop during exercise with compressed air did not benefit from supplemental oxygen. Nonetheless, the degree of dyspnea during exercise with compressed air could not be used to predict the beneficial effect of supplemental oxygen.