Abstract
When two successively presented letters are visually identical (for example, AA) rather than identical in name only (for example, aA) subjects can more quickly name the second letter and can more quickly decide if the two letters have the same name. For both tasks the maximum interstimulus interval (ISI) over which this visual priming effect is observed depends upon the relative spatial positioning of the two stimuli. When they are superimposed the effect can survive an ISI of between 6 and 15 s, but when they appear in increasingly different spatial locations the maximum ISI is reduced to between 0.5 and I. O. S. With successive letters spatially separated in the naming task the visual priming effect is replaced by a comparable name priming effect as the ISI is extended beyond 0.5-1.0 s. Previous studies of letter matching (for example, Posner and Keele, 1967) have underestimated the durability of visual memory, a fact that is difficult to reconcile with the “levels of analysis” approach to visual memory.