Hypothesizing that severely language delayed children lack sufficient understanding of causal structure to comprehend stories, a reading program concentrated on developing the causal reasoning of 10 children between the ages of 10 and 12 with a verbal intelligence quotient two to three years below average. Instructional activities in the eight-week program stressed (1) identifying story information categories and causal relationships in simple and more complex stories, (2) completing macro-cloze exercises, (3) creating story trees or graphs, and (4) initiating and responding to inferential questions about causal relationships in stories. Comparing the instructional group's pretest and posttest causal reasoning and recall scores with those of six students of similar age and aptitude indicated that the instructional group did make progress but that the progress varied across tasks and across children. The results showed no significant improvement in student recall. The data demonstrated the importance of matching the instruction to children's readiness level and indicated that global effects of new instruction are unlikely. It also indicated that causal relationships among elements within and between story episodes are important aspects of reading skill. (Extensive tables are appended.) (MM)