A single novel word among several familiarized words may be localized more effectively than the familiarized words (novel popout). Early demonstrations of novel popout attributed the effect to the capture of attentional resources by novel stimuli. Klein (1995, 1996)argued that differential recollection of novel versus familiar words could alternatively account for the popout effect. In the present experiments, participants judged which of four locations contained a physically brighter word. A bright novel word was localized significantly better than a bright familiar word in one-novel/threefamiliar arrays, inconsistent with a retrievability account of novel popout. However, a bright familiar word was also localized better than a bright novel word in three-novel/one-familiar arrays, inconsistent with the mismatch theory proposed by Johnston and Hawley (1994). The results suggest that familiarity and novelty provide a perceptual segregation of the odd item; superior brightness discrimination at that location may be due either to attentional capture or to locational ambiguity within the larger group.Properties that distinguish one object from an array of others, such as luminance, color, or abrupt onset, may capture attention in a stimulus-driven fashion-that is, independent of a person's strategies or intentions (Theeuwes, 1991(Theeuwes, , 1992Yantis & Jonides, 1984; but see Jonides & Yantis, 1988). Research by Johnston and colleagues (Hawley, Johnston, & Farnham, 1994;Johnston, Hawley, & Farnham, 1993;Johnston, Hawley, Plewe, Elliott, & DeWitt, 1990) suggested that the novelty of an object may similarly capture attention, causing it to perceptually pop out against an array of familiarized items.The novel popout effect has typically been demonstrated in a task in which participants are shown an array of four different words followed by a single probe word; they then attempt to indicate the initial location of the probed word. Some of the words are familiar in the context of the experiment because they appear repeatedly across trials; other words are novel, appearing only once. If an array contains three familiar words and one novel word, localization of the novel word is likely to be enhanced relative to an all-novel array (between-arrays novel popout), and localization of a familiar word may suffer relative to an all-familiar array (familiar sink-in). In some experiments, the novel word in a one-novel array is localized better than the familiar words in that array (withinarray novel popout), despite higher performance on allfamiliar arrays than on all-novel arrays (baseline effect).This work was funded by a Faculty Research Grant awarded to the second author. We thank Molly Brown, Michelle Gruhn, Lauren May, and Tammy Schillinger for their assistance with data collection. We are also indebted to Ray Klein, Bradley Gibson, William Johnston, and Jan Theeuwes for helpful suggestions and comments on earlier versions of this manuscript. K. A. Diliberto is now at Berry College, Mount Berry, GA. Correspondence concerni...