A long-held belief is that human attention can be deployed voluntarily according to observers’ goals (top-down) or shifted automatically to the most salience object in the environment (bottom-up). Recent studies suggest a third category of attentional control: selection history. By this view, an observer’s experience in performing a task that requires the control of attention could automatically affect subsequent attention deployment in the task. This thesis examined selection history mechanisms of attentional control in two visual search phenomena. The first phenomenon is known as the Contextual Cueing Effect (CCE), and refers to an increased search efficiency when a specific distractor configuration is repeatedly associated with a specific target location (Chun and Jiang, 1998). In one study, we found a CCE when one repeated configuration was associated with up to four different target locations, suggesting that the CCE may involve mechanisms other than attentional guidance by one-to-one context-target associations. In another study, we found that the CCE was not affected by concurrent working memory load, and that there was little correlation between the magnitude of the CCE and working memory task performance when measured separately in the same participants. These results suggest that working memory may not be involved in such contextual learning. The second phenomenon is known as the the Item-Specific Proportion Congruent (ISPC) Effect, and refers to item-specific learning that controls the extent to which salient distractors capture attention. Through manual response and eye movement measures, we demonstrate that the ISPC effect reflects the search process itself, rather than processes that precede or follow search. We propose does item-specific learning produces transient changes in the activation of goal-related processes that mediate attention capture. / Thesis / Doctor of Psychology (PsyD) / Where we attend in visual space can be affected involuntarily by memories of how we have attended to visual space in the past. In other words, automatically retrieved memories can control our visual attention independent of volition. This thesis examines two visual search phenomena that display this type of memory-based control over attention. The first phenomenon reveals that search performance improves with experience searching through the same set of visual distractors on multiple occasions. We demonstrate that this form of learning is remarkably flexible; it can occur for multiple targets associated with the same set of distractors. We also demonstrate that this form of learning probably involves long-term rather than short-term memory mechanisms. The second phenomenon reveals how memory-based processes can prevent attention from being captured by a salient distractor. Eye movement data reveal that this form of learning impacts search itself, rather the processes that precede or follow search.