Sperling (1960) reported a classic study of visual attention. In the whole-report condition he presented subjects with brief presentations (50 msec, followed by a mask) of visual arrays of letters (3 rows and 4 columns) and found that on average they could report back 4.4 letters. In the partial-report condition he gave subjects an auditory cue to identify which row they would have to report. Then he found that they were able to report 3.3 letters in that row. As he delayed the presentation of the auditory cue to 1 second after the visual presentation he found that subjects' recall fell to about 1.5 letters. Since subjects' recall at a second's delay fell to about a third of the whole report level, the obvious interpretation was that they were able to report as many items from the cued row as they happened to encode without the cue. This research has been interpreted as indicating that subjects have access to all the letters in a visual buffer but they have difficulty in reporting them before they decay away.