To assess the feasibility of using odors as a potential mechanism for treating sleep apnea, we set out to test the hypothesis that odorants delivered during sleep would modify respiratory patterns without inducing arousal or wake in healthy sleepers. We used 2 mildly trigeminal odorants: the pleasant lavender and unpleasant vetiver oil and 2 pure olfactory odorants: the pleasant vanillin and unpleasant ammonium sulfide. During sleep, an olfactometer delivered a transient odorant every 9, 12, or 15 min (randomized), providing 21-37 odorant presentations per night. Each of 36 participants was studied for 1 night and with 1 of the 4 different odorants tested. In addition to standard overnight polysomnography, we employed highly accurate measurements of nasal and oral respiration. Odorants did not increase the frequency of arousals or wake but did influence respiration. Specifically, all 4 odorants transiently decreased inhalation and increased exhalation for up to 6 breaths following odor onset. This effect persisted regardless of odorant valence or stage of sleep. These results suggest that the olfactory system may provide a path to manipulate respiration in sleep.