Learning often involves integration of information from multiple texts. The aim of the current study was to determine whether relevant information from previously read texts is spontaneously activated during reading, allowing for integration between texts (experiment 1 and 2), and whether this process is related to the representation of the texts (experiment 2). In both experiments, texts with inconsistent target sentences were preceded by texts that either did or did not contain explanations that resolved the inconsistencies. In experiment 1, the reading times of the target sentences introducing inconsistencies were faster if the preceding text contained an explanation for the inconsistency than if it did not. This result demonstrates that relevant information from a prior text is spontaneously activated when the target sentence is read. In experiment 2 free recall was used to gain insight into the representation after reading. The reading time results for experiment 2 replicated the reading time results for experiment 1. However, the effects on reading times did not translate to measurable differences in text representations after reading. This research extends our knowledge about the processes involved in multiple text comprehension: Prior text information is spontaneously activated during reading, thereby enabling integration between different texts.