Ten trained, audiometrically normal young adults heard a 1K-Hz reference tone at 50-dB or 80-dB SPL. They were instructed to maintain the reference loudness throughout a three-minute 1K-Hz fixed-frequency run. Seven test signals, consisting of one sustained and six differentially interrupted pure tones, were employed for loudness-memory tracking. Preliminary training eliminated learning effects. Thirty-second pretracking adaptation periods produced loudness adaptation which was complete to asymptote for each test signal. The tracking levels were inversely related to the signal’s duty cycle and were independent of other temporal parameters such as on-duration and interruption rate. The occurrence of the Type V Bekesy audiogram is attributed to the differential effects of memory upon the loudness of sustained and interrupted pure tones.