These experiments examined the ability of musically trained subjects to hear out individual partials in complex tones with partials uniformly spaced on a scale related to the equivalent rectangular bandwidth (ERB) of the auditory filter. ERB spacings of 0.75, 1.0, 1.25, 1.5, and 2 were used, and the central component always had a frequency of 1000 Hz. All components had a level of 65 dB SPL. On each trial, subjects heard a pure tone (the "probe") followed by a complex tone. The probe was close in frequency to one of the partials in the complex, but was mistuned downward by 4.5% on half the trials (at random) and mistuned upward by 4.5% on the other half. The task of the subject was to indicate whether the probe was higher or lower in frequency than the nearest partial in the complex. The partial that was "probed" varied randomly from trial to trial. Scores for the highest and lowest components in the complexes were generally high (> 90%) for component spacings greater than 1 ERB, but worsened somewhat for ERB spacings of 0.75 and 1.0. Scores for the inner components were close to chance level at 0.75-ERB spacing, and improved progressively as the ERB spacing was increased from 1 to 2 ERBs. For ERB spacings of 1.25 or less, the scores did not change smoothly with component frequency; marked irregularities were observed, as well as systematic errors. An explanation for these is suggested in terms of irregularities in the middle ear transfer function. Performance for the inner components tended to be worse for component frequencies above 1000 Hz than below 1000 Hz. It is suggested that this happens because the pitches of partials are partly coded in the time patterns of neural impulses (phase locking), and the precision of phase locking deteriorates progressively with increasing frequency above 1000 Hz. The auditory filter shapes of the subjects were measured for a center frequency of 1000 Hz, using the notched noise method. One subject had a broader auditory filter than the other three subjects, and this same subject generally had more difficulty in hearing out partials from complex tones than the other subjects.