Imagine that you are correcting a manuscript for typos. Spotting an erroneous blank space is easy if the additional blank is located between letters within a word. However, it is fairly difficult to find if it is located between words. One factor that contributes to this asymmetry in visual search difficulty is singleton capture. Many studies have confirmed that visual stimuli that differ by one or several features from their more homogeneous neighbors capture attention (for reviews, see Burnham, 2007;Theeuwes, 1992Theeuwes, , 2004. Van Zoest, Donk, and Theeuwes (2004), for example, found that a right-tilted bar among several vertical bars captured attention. Accordingly, in our example above, a blank space between two flanking letters within a word captures attention more easily and can be spotted better than, for example, a blank space flanked by another blank space on one side and a letter on the other.According to Bacon and Egeth (1994), attentional capture by a singleton is goal dependent. Only if a participant is set to search for a singleton will a singleton capture attention. These authors observed that for attentional capture by an irrelevant color singleton to occur, participants had to search for a shape singleton. If one green circle among several green diamonds had to be searched for as a (shape) singleton (because it contained the response-relevant target line), a red diamond as an irrelevant color singleton captured attention. However, if several shape singletons (e.g., a circle, a triangle, and an octagon) were presented at the same time, participants had to change search mode: To find the target by searching for a (shape) singleton was no longer sufficient; rather, participants now had to search for a specific shape (e.g., a circle). In line with a goaldependent effect, under these conditions, the irrelevant color singleton failed to capture attention.By contrast, Theeuwes (1992Theeuwes ( , 2004 and others (e.g., Burnham & Neely, 2008) have regarded singleton capture as being stimulus driven in nature. On this view, any strong local feature difference has the power to attract attention (Bergen & Julesz, 1983;Burnham & Neely, 2008). This position fits well with results from eyetracking studies: Participants freely viewing 2-D images of natural scenes tend to fixate on strong local feature differences in color, luminance, or orientation (see Itti & Koch, 2001;Parkhurst, Law, & Niebur, 2002).In the present study we tested one prediction of computational theories concerning the role of awareness for singleton capture. Computational theories consider singleton capture to be a consequence of objective feature differences: Typically, an image location's potential to attract or capture attention is assessed as its pooled or summed standard deviation across several visual dimensions (e.g., color, orientation, luminance). This aggregated feature difference is then "compared" with similarly calculated standard deviations at other image locations (see Itti & Koch, 2001;Parkhurst et al., 2002 London, Englan...