A technique obtaining a precise mapping of interactive effects in the pitch memory store is described. Subjects were required to compare two tones for pitch when these were separated by a 5-second interval during which six other tones were played. In the second serial position of the intervening sequence there was placed a tone whose pitch bore a critical relationship to the pitch of the first test tone. When the critical intervening tone was identical in pitch to the first test tone, memory facilitation was produced. As the separation in pitch between these two tones increased, errors rose progressively, peaked at a separation of 2/3 tone, and declined roughly to baseline at a whole tone separation. It is concluded that the pitch memory store is arranged logarithmically in a highly ordered and specific fashion.