Distractor interference is subject to dilution from other nontarget elements, and the level of dilution is affected by attention. This study explores the nature of dilution when the location and color of the target is known in advance. Experiments 1 and 2 show that attention is effectively limited to the precued region, so that it is the nontarget letters appearing at the cued locations that are responsible for most of the dilution, and not those appearing at the uncued locations. Furthermore, this dilution occurs relatively early in processing. Experiment 3 demonstrates that top-down attentional control can prevent dilution, because foreknowledge of the target color leads to quick attention shifts. Experiment 4 illustrates bottom-up attentional control in preventing dilution when the distractor is a color singleton that is segregated from the diluting nontargets. The results show that dilution is modulated by both top-down and bottom-up factors, that it can occur even when attention is restricted to a relatively small region, and that it occurs early in processing, but not so early that it avoids the effects of attention. They provide new challenges for earlier accounts suggesting that dilution is widespread and unfettered by attention. Likewise, some parts of the results are difficult to reconcile with the alternative perceptual load theory, but they do support a form of dilution that is limited by attentional boundaries. Because of that link to attention, dilution is a useful tool for measuring how attention is guided by information about target location and color.