Adults who have completed less education live shorter and less healthy lives. All indicators are that this disparity is growing. Educational outcomes are closely tied to literacy and reading skill. In order to improve educational outcomes for all students we need a research base that reflects the population. This is currently not the case. Still today most research on reading is based on English-speaking monolinguals. Yet, we know from recent research that reading for bilinguals is fundamentally different from monolinguals. Therefore, theories based on monolingual speakers cannot be applied across the gamut of language experience and linguistic diversity. The long- term objective of this line of research is to test and extend current theories of reading so that they address the unique cognitive characteristics of bilingual readers. The short-term goal of this proposal is to test and extend a specific, influential model of discourse comprehension, the landscape model, to bilingual readers. According to the model readers integrate information across text through two types of retrieval: cohort-based retrieval and coherence-based retrieval. Cohort-based retrieval is an implicit memory process, based on an automatic spread of activation. Depending on a reader’s goals for comprehension, they may engage in more effortful, coherence-based retrieval. This form of retrieval is controlled, deliberate and engaged when a discrepancy is detected. Through this form of retrieval a reader seeks to establish global coherence, seeking a way to integrate incoming information with an on-going representation of the text as a whole. The specific aims of this proposal are to examine how the strength of cohort-activation (Specific Aim 1) and coherence-based retrieval (Specific Aim 2) are affected by the (1) match in language across texts, (2) genre of text and (3) cross-language reading experience for bilinguals. The overall approach to achieving these aims will consist of presenting Spanish-English bilingual university students pairs of text passages to read while their eye-movements are recorded. Half of the passages will be narratives and half will be expository, science texts. Passages will contain some facts that are revised in the second passage, requiring the reader to update their representation of the discourse. Second passages will either be in the same or different language as the first passage. The impact of these variables on reading times for critical areas of text and responses to follow-up comprehension questions will be examined through linear mixed-effect models and ANCOVA’s. We will also include in the analyses scores that reflect participants’ experience reading in each of their languages.