Effect of absolute and relative gap sizes in visual search for closure

In three visual search experiments using completely closed squares and squares with gaps, the physical gap size and the ratio of gap size relative to contour length were systematically manipulated. In search for an open square among closed squares, the search rates were fast and independent of the physical gap size or the gap-to-contour ratio. This is consistent with previous studies (e.g., Treisman & Souther, 1985). In contrast, in search for a closed square among open squares, the gap-to-contour ratio determined the search rates, irrespective of physical gap size. The latter result suggests that visual search is based not on local contour information but on global closure information, which seems to be constructed at some higher level of visual processing, probably at the level of perceptual representations of shapes and objects (e.g., Elder & Zucker, 1993; Rensink & Enns, 1995).