50 healty normosmic students obtained odor thresholds for butanol using the Sniffin` Stickstest (SnS), as well as the method of limits (ML) in ascending and descending order. The procedureswere repeated about 2 weeks later. For each session and method the performancetime was taken and subjects rated the pleasantness of the odor and the strain required by thetask using a category anchored Visual Analogue Scale. For all subjects plausible thresholdscould be obtained by the SnS procedure and by the ascending ML. For 22 out of 50 subjects,however, thresholds could not be obtained by the descending ML. Further analysis of the datarevealed worse performance and significantly lower retest-reliabilities for those subjects inthe ascending ML and the SnS-procedure as well.