Perception takes time, and the amount of time required to perceive one attribute of a given sound as accurately as possible may far exceed the duration of the stimulus itself. Such a view has been supported by a number of auditory discrimination experiments in which performance was assessed as a function of the time interval separating the two stimuli presented on a given trial. When short stimuli were used, it was typically found that discrimination performance improved as the interstimulus interval (ISI) increased, in a range with an upper boundary that was often well above 200 or 300 msec. Tanner (1961) may have been the first to report such data. In an experiment on intensity discrimination, he found that performance (measured in terms of d′) was optimal for an ISI of about 600 msec, decreased slowly for larger ISIs, and dropped much more abruptly for smaller ISIs. In his brief description of this experiment, Tanner did not specify his stimuli. However, his basic result was replicated by Sorkin (1966), who used 50-msec tone bursts in a similar experiment and also found an optimal ISI of about 600 msec when the two stimuli produced on a given trial were presented to the same ear. Sorkin further observed that the optimal ISI was only slightly smaller if the two stimuli were instead presented to opposite ears. This observation is important because it rules out the idea that the need for an appreciable ISI for optimal performance could be due entirely to a forward masking of the second stimulus by the first one at very short ISIs. Forward masking was minimized when the two stimuli were presented to opposite ears. Thus, at least in this case, the ISI effect clearly meant that several hundreds of milliseconds were needed to perceive the first stimulus optimally after its offset. Taylor and Smith (1975) and Berliner, Durlach, and Braida (1977) corroborated Tanner's (1961) and Sorkin's (1966) findings in subsequent studies on intensity discrimination. Furthermore, analogous results were obtained for other acoustic continua. For instance, the studies of Small and Campbell (1962) and Carbotte (1973) indicate that it is necessary to use an ISI exceeding 200 msec in order to measure optimal performances in the discrimination of short sounds differing in duration. Several hundreds of milliseconds also seem to be needed for the optimal discrimination of formant transitions in synthetic syllables, according to Pisoni (1973). Finally, in the domain of pitch, an effect of the same type was reported by Massaro and Idson (1977). Their stimuli were 20-msec bursts of sinusoids differing in frequency and, as in Sorkin's study, the two stimuli used on a given trial were presented either to the same ear or to opposite ears. In both cases, discrimination performance, measured in terms of P(C) (percent correct), appeared to improve from a level close to chance for an ISI of 5 msec to an optimum level for an ISI of about 350 msec.Before the study that we just mentioned, Massaro had conducted a series of related experiments in which th...