Past research has shown a tendency for readers who are given multiple perspectives on a domain to excel in solving problems with material from that domain. Mannes (1994) hypothesized the reinstatement-and-integration strategy as a possible processing account of these findings. According to this strategy, when a reader encounters a topic that has been learned about previously, the proposition representing that concept becomes activated in memory, as do propositions representing some information about the original learning context in which it was encountered. When the original and current contexts differ, relational processing takes place, affording the reader a more interconnected memory representation of the domain under study. In this article, three experiments are reported that provide on-line evidence for the reinstatement-and-integration strategy. Subjects read a target text at a self-determined rate after studying an outline that either did or did not conform to the perspective of the target text. Those readers for whom the text and outline presented multiple perspectives took longer than did readers for whom the experimental materials presented a single perspective. A process explanation in terms of the reinstatement-and-integration strategy is presented.