Thirty-two Ss between 22 and 59 years of age yielded detection thresholds for 4 odorants over 4 sessions. The thresholds decreased and reliability increased over the course of testing. High intercorrelations between odorants and the stability of an S's relative position within the threshold distributions showed that a general factor of sensitivity dominated the outcome. Age contributed strongly to intersubject variation. Even among these nonelderly individuals, it accounted for up to 2 orders of magnitude in threshold performance. Other important factors included superiority of the right nostril and a negative correlation between the mean and variance of threshold distributions. Scant attention to the correlation may have contributed to overestimation of the frequency and specificity of specific anosmia. A clinically relevant outcome was that measurement of threshold for diagnostic purposes can generally rely on just 1 odorant.