Visual marking beside the mark: Prioritizing selection by abrupt onsets

In a standard visual marking experiment, observers are presented with a dis play containing one set of elements (old elements) followed after a certain time interval by a second set of elements (new elements). The task of obse rvers is to search for a target among the new elements. Topically, the time to find the target depends only on the number of new elements in the displ ay and not on the number of old elements, showing that observers search onl y among the new elements. This effect of prioritizing new elements over old elements is explained in terms of top-down inhibition of old objects-that is, visual marking (Watson & Humphreys, 1997). The present study addressed whether this prioritizing is in fact mediated by top-down inhibition of old objects, as suggested by Watson and Humphreys (1997), or whether it is med iated by the abrupt onsets of the newly presented elements (Yantis & Jonide s, 1984). In three experiments, the presentations of the old and new elemen ts were or were not accompanied by a luminance change. The results showed t hat if new elements were equiluminant with the background, no visual markin g occurred, suggesting that new elements must have a luminance onset in ord er to be prioritized over old elements. Implications for cur-rent theories on visual selection are discussed.