Olfactory receptor neurons (ORNs) in the nasal cavity sense VOCs. These ORNs undergo constant turnover and shift in their response as a function of fatigue and exposure. Despite this variability, olfactory perception appears constant over time, suggesting robust mechanisms compensating for fluctuations. Here we put forth a novel hypothesis where the two-nostril system allows maintaining constancy despite drift in the olfactory image. We propose a winner-takes-all competition between nostrils in pattern identification. If one nostril registers strong identification, the other recalibrates its image accordingly. To test this hypothesis, we constructed an apparatus allowing delivery of different stimuli to each nostril. The apparatus contains a barrier between nostrils, 3D-printed to fit each participant’s nose. In an ongoing experiment, participants were trained to identify four complex odorants (O1-4). After participants reach criteria, in an altered-identification task, we incrementally replace O1 with a new, unidentified odorant (OX), one nostril at a time, while keeping O2-4 the same. Separate participants completed the same tasks without nostril separation to control for learning. After completely replacing O1 with OX, participants named the odors. Results from 26 participants (N=13 in each group) revealed a significant effect of the manipulation: In 52% of trials, participants named OX as if it were O1. Participants named OX correctly in only 19% of trials (z =-14.4, p<.01). Despite a soft trend in the direction of our hypothesis whereby this perceptual shift was greater in the group with nostril separation (Separated: 54% shifted, only 15% correctly named, Without: 50% shifted, 23% correctly named), this interaction was not significant (OX: χ²=.41,p=.52, O1: χ²=.01,p=.93). The current data imply a remarkably fast shift in olfactory perception, but only minimally support our hypothesis. Final determination on this front awaits completion of the intended cohort.