Four human observers were run in a two‐alternative, forced choice procedure to measure the detectability of a 500‐Hz tone in the presence of each of 25 digitally generated, 100–3000 Hz maskers. The masker was randomly selected on each trial within blocks of 80 trials. The average percentage of correct responses across all maskers was within the range usually reported. The percentages for the individual maskers varied from chance to near perfect, however. The three maskers which led to the best detection and the three which led to the worst detection were used in the second portion of the study to stimulate low‐frequency, auditory‐nerve fibers of the chinchilla. Fifty presentations of each signal and masker combination were used to obtain spike rates and phase‐locking measures. These results are compared with those obtained from the first portion of the study and with measures of the physical characteristics of the stimuli. [Supported by NIH Grant Nos. NS1928 and NS16396.]