Four experiments were concerned with priming effects that occur when participants attempt to identify each letter in a briefly presented 4-Ietter display. At each position of the display, 1 alternative (the target) was more probable than 4 alternatives (nontargets). The basic finding, which held true over several manipulations of display characteristics, was that retroactive effects were generally much stronger than proactive effects. These strong retroactive effects contrast with prior findings that (a) proactive effects usually dominate and (b) repetition yields inhibition (repetition blindness). This asymmetry was also noted in an additional experiment in which participants responded to only 1 letter in the array.Responses associated with stimuli presented in succession can interact in two basic ways. The first stimulus in a pair can affect a succeeding stimulus, giving proactive or forward priming, or the succeeding stimulus can affect the preceding stimulus, giving retroactive or backward priming. Normally, proactive effects are much stronger than retroactive effects. Meyer and Schvaneveldt (1971) provide an often-cited instance of proactive priming. They presented two stimuli simultaneously, one above the other. The stimulus on the top was a word, and the stimulus on the bottom was either a related word, an unrelated word, or a nonword. The participant's task was to judge, as quickly as possible, whether the second stimulus was a word or nonword. The finding that reaction times were faster for related word pairs than for unrelated word pairs is commonly interpreted in terms of spreading activation (