Dyspnea in patients with advanced cancer is a common symptom that is difficult to treat. This study investigated whether oxygen helps to relieve rest dyspnea in patients with advanced cancer. In a single-blind controlled trial, oxygen and air were administered in random order to hospice patients reporting dyspnea at rest. Measurements of arterial oxygen saturation, lung function, and dyspnea (using a visual analogue scale [VAS] and Borg score) were made before and after each gas had been given for 15 min. Data from 38 patients were used: analysis of variance revealed that mean VAS levels during baseline conditions, breathing room air (59 mm), were significantly reduced after administration of either air (48 mm; p < 0.001) or oxygen (45 mm; p < 0.001); there was no significant difference for the mean VAS scores between oxygen and air administration. There was no statistically significant order of treatment effect. There was no difference in the response to oxygen or air in patients with a history of cardiopulmonary disease. The improvement in dyspnea with oxygen could not be predicted from a subject's initial level of hypoxia. Results suggested that benzodiazepines may potentiate the effect of oxygen. The overall conclusion is that oxygen and air can have a significant effect in reducing dyspnea at rest in patients with advanced cancer.