Three experiments examined the flexibility of attention to the color and location dimensions of visual stimuli. Displays typically consisted of two curved lines, one a target and the other a distractor. Subjects were precued to the location of the target, and they made a speeded response to that item's direction of curvature. In Experiment 1, a color difference between the stimuli facilitated selective attention to targets. Experiment 2 demonstrated that this effect was not automatic. Although component color differences were necessary for selective attention, they were not sufficient unless they reliably specified the target. Furthermore, when color was informative but the target was in an unexpected color, performance was markedly impaired. A third experiment, using a response-deadline technique, showed that stimulus color was resolved before stimulus location and that attention can be exclusive to the color dimension. The results were interpreted in terms of a model assuming that stimulus color and location are processed independently. Subjects attend to the dimension resolved most quickly which can reliably specify the target.