developed the now classic cuing technique for manipulating spatial attention independently of eye movements. In this paradigm, participants typically have to detect or discriminate a target and respond manually by pressing the appropriate response key. The target stimulus is preceded by a cue that provides information about the location of the upcoming target. In a cuing version typically referred to as endogenous, a centrally displayed arrow points to the likely target location. Participants are instructed to use the arrow and focus their attention on the cued location before the appearance of the target. Results show that processing of the target is enhanced at the cued location. In the so-called exogenous version of the cuing paradigm, before the appearance of the target, the participant's attention is pulled to one of the locations by an uninformative peripheral cue (usually an abrupt increment or decrement in luminance). The target then appears either at the cued or the uncued location. Shortly after presentation of the cue, response times (RTs) for targets at the cued location (valid cue) are fast and accuracy is high, relative to when the target appears at the uncued location (invalid cue). However, when the interval between the cue and the target is long, RTs to targets at the cued location are delayed in comparison with those to targets at the uncued location (i.e., inhibition of return; Posner & Cohen, 1984).These cuing experiments have been important for understanding spatial attention as a "spotlight." Indeed, Posner et al. (1980) described attention as a "spotlight that enhances the efficiency of detection of events within its beam" (p. 172). Over the last 25 years, the spotlight metaphor has generated several important research questions regarding how the beam moves through space, whether the beam can split, and the extent to which it can be highly focused or not (for a review and discussion, see Cave & Bichot, 1999). Alternative approaches do not consider RT benefits and costs obtained in typical cuing tasks to be the result of a moving "spotlight," but instead to result from an efficient allocation of attention over the visual field. For example, Shaw and Shaw (1977;Shaw, 1978) proposed a parallel model of attention suggesting that attention can be allocated flexibly to multiple locations in parallel. According to their model, the attentional system has a fixed capacity for visual processing that is optimally distributed over the different spatial locations. The time it takes for a stimulus to be identified is a function of the capacity allocated to the corresponding location. Attention is divided across a visual scene on the basis of the likelihood of the target location. When chances are high that the target will be presented at a certain location, more attentional resources will be allocated to that location. Therefore, in an endogenous cuing experiment, more resources will be allocated to the valid location than to the invalid location. Because of the higher amount of resources allocated to t...