BACKGROUND: In visual search for a target item among distractors, attention can be guided to a target by basic features of that target (Find the RED "x".) Chun and Jiang's (1998) "contextual cueing" effect shows that RTs are speeded if the spatial configuration of items in a scene is repeated over time: If the items are laid out in this pattern, then the target is at location XY. We ask if featural properties of the scene as a whole guide attention? (i.e. If the display is mostly RED, then the target is at location XY). METHOD: We ran a series of visual search tasks where the dominant color or orientation, present in the background, predicted the exact target location. Unlike guided search, the target did not have a defining color or orientation texture. Unlike contextual cueing, the spatial layouts of items were not informative. The basic feature could serve as an endogenous cue to target location. In Exp. 1, the mean color or texture of the background could predict the exact target location while distractor locations were randomized. In Exp. 2, the background was presented prior to the search display, to see if cueing observers ahead of time aided performance. In Exp. 3, the spatial configuration of the stimuli was the same for all trials while the features of the background cued target location. RESULTS: In Exp 1, there was no benefit in search when mean features of the background predicted target location in the absence of an exact target-distractor configuration. In Exp 2., a preview of the background of up to 800 msec, failed to improve search. However, in Exp. 3, when the configuration of the stimuli was invariant, predictive features in the background produced a benefit. Endogenous cueing by background features like color or texture is possible but it appears to be easily vetoed by other background information such as spatial layout.