In three half���hour practice sessions over three days, 88 grade 5 students from a rural city in Australia were given practice in one of the following pre���reading exercises: (1) an irrelevant task (control); (2) a pictorial introduction to the topic; (3) content���directed questions to be answered; (4) free���association to the passage's titles. On day 4, post���test measures were taken on literal and inferential comprehension using the practiced pre���reading activity. In a later session, students were asked to rate the efficacy of each type of pre���reading activity and to provide rationales for their first preferences. While the pictorial introduction group performed best overall on comprehension, it was found (i) that any topic relevant introduction significantly enhanced background���only inferential com��� prehension, (ii) pictorial introductions facilitated inferences requiring the linkage of disparate textual elements, and (iii) directed questioning significantly aided literal comprehension. It was also found that students favored pictures as introductions, and explained their preferences mostly in terms of the provision of relevant information and the initiation of relevant cognitive���processing operations. �� 1987 Taylor & Francis Group, LLC.