In an experiment designed to investigate the time decay of auditory stream biasing (ASB), subjects were required to listen to a 10-sec induction sequence of repeated tones (AAAA...) designed to bias the listener's percept toward hearing an A stream. The induction sequence was followed immediately by a silent interval (0-8 sec), and then a short ABAB ... test sequence. To measure the amount of ASB remaining at the end of the silent interval, subjects were asked to indicate whether the test sequence was temporally coherent or had segregated into separate A and B streams. A plot of the mean number of segregation responses against silent-interval duration indicated that the overall time decay of ASB can be ascribed by an exponential decay function with a time constant of tau = 3.84 sec, with musicians having a longer time constant (tau = 7.84 sec) than nonmusicians (tau = 1.42 sec). The length of the time constants for musicians and nonmusicians suggests that the mechanism responsible for ASB is associated with long auditory storage and that future experiments investigating auditory streaming phenomena should use interstimulus intervals of at least 8 sec.