Visual search, or the ability to locate a single target from a field of distractors, is used as a means of assessing the capacity of the visual system with regard to the detection of low-level stimulus dimensions such as color, line orientation, and other specific perceptual features and higher level characteristics such as three-dimensional (3-D) orientation implied by pictorial cues to depth or occlusion. Target-distractor differences of importance to the visual system result in rapid (and possibly preattentive) target location while differences of lesser importance (or differences that are less salient) tend to elicit slower, less efficient search performance (search in this latter case may require attentional resources.) Adult abilities in this type of task have been tested extensively, but comparatively little is known about the development of these abilities. The present investigation was conducted to test, across a range of ages in childhood, the ability to detect visual search targets defined by a low-level critical feature (a texton; B. Julesz, 1981) and by static (pictorial) cues to depth. Children 2 to 9 years of age and adults were tested with each type of target. Previous data from infant studies predicted that the texton-defined differences would elicit highly efficient search performance, and this proved to be the case. When no texton difference was present, however, children's search was highly inefficient. Children between 4 to 9 years of age also proved to be quite efficient at locating a target, differing only by a pictorially rendered 3-D orientation from the distractors, but were much less efficient when the static cues to 3-D orientation were removed. These findings confirm previous reports that the basic processes underlying visual search are functional early in development, but suggest that children's search is less efficient than adult search when attention is required for object localization.