A study examined the emergent patterns and anomalies of adult readers as they unload information from their long term memory into their retellings of text. Participants were 16 graduate students enrolled in a Reading Education Seminar at Indiana University. Participants formed a heterogeneous group which included speakers of six different languages and whose command of spoken English varied widely. The data were collected as blind-coded verbatim transcriptions of the recorded oral retellings. After interpreting and classifying the content independently, a tentative taxonomy was established in concert which was derived from the observed content of the retellings. All transcriptions were analyzed independently by four researchers for the following elements: generalization, thesis statement, major points, supporting details, supplementations, coherence, completeness, and comprehensibility. The blind transcriptions were identified by name and their content and rating were reviewed in relation to the known facts about nationality, educational and cultural background, familiarity with the text selected, and whether the participants spoke English as a native, a second, or a foreign language. Results included three major findings: (1) although participants came from various cultural and educational backgrounds, a majority of them repeatedly mentioned and expanded upon certain key points from the text; (2) many readers who expressed familiarity with or interest in the topic prior to reading the passage showed a higher incidence of reader perspective; and (3) the richness of the retelling affected the overall rating. (Two appendixes containing a data summary and taxonomy are attached.) (MG)