The amount and structure of prose learning derived from vocabulary definition and graphic post-organizer construction methods of independent study tasks is examined. Undergraduate subjects (n=144) read a 2000-word passage in a 20-minute reading period. Two subject groups performed text information tasks while a third acted as a control to measure recall derived from a reading-only strategy. Two days later, all subjects, including an additional control group (n=44) who had neither read nor studied the material, completed a multiple-choice test on the passage. Results of a multivariate analysis of variance indicated that both study procedures were successful, but that the two groups differed on specific types of information learned. A single discriminant function was found to separate the two study groups. Vocabulary rehearsal led to the learning of greater amounts of subordinate level information, while graphic post-organizer usage led to the learning of more superordinate information. This study was limited due to low internal consistencies, short test drawn from a single content area, and learning measured on the basis of recognition only. (Author/PN)