Ten participants sought to detect four odorants: benzaldehyde, pyridine, and two alcohols, n-butyl and n-amyl alcohol, that smelled similar to each other. These were presented repeatedly on 3 successive days. The sequence of testing during a session made it possible to determine whether experience with one odorant would specifically facilitate the detectability of a similar-smelling odorant and whether any such facilitation would restrict itself to the nostril through which the experience was gained. Neither of these possibilities occurred. Instead, measured sensitivity increased rather uniformly both within and across days. Net gain from beginning to end exceeded an order of magnitude. Averaging across sessions gave a picture of smaller than usual individual differences, under 20 to 1, attributable mainly to general rather than odorant-specific differences in sensitivity. The results indicate that thresholds gathered in customary brief testing will underestimate olfactory sensitivity and overestimate individual differences. Incorporation of a reference odorant into threshold experiments should increase comparability among studies.Olfactory sensitivity as typically measured in the laboratory varies markedly from one person to another (Punter, 1983). Based on his own extensive data, Amoore (1980) developed the rule of thumb that about 96% of the population falls between lA6 and 16 times the mean threshold for an odorant, that is, within a 256-fold range.Does such high variation reflect only individual differences in absolute sensitivity? Substantial extraneous variation could occur from poor stimulus control, unsophisticated psychophysical methodology, or collection of insufficient data. Studies of olfactory threshold have often relied, for instance, on a single brief determination of threshold per participant. Even with more thorough measurements, however, reliability has received scant attention. Hence, we know little about whether variability between people commonly exceeds variability within a person tested at different times.Observers may differ in sensitivity in two ways: (1) One' observer may prove more sensitive than another to virtually all odorants. This situation could reflect the operation of some general factor, including something as simple as how many molecules can gain access to the receptors within the nasal cavity. (2) An observer may prove more sensitive than average to some odors and less sensitive to others. If observers show idiosyncratic yet stable disparities, then the possible explanations could prove numerous.The present study addressed the issue of individual variation and reliability, using standard threshold methodology to gather observations on four odorants for the same individuals both within a day and across days. The testing thereby allowed a glimpse of whether participants would show systematic changes both over time and across odorants. The sequence of testing within a day was designed to permit conclusions about the possible mechanism and specificity of any gains in performanc...