Recent research has shown that when individuals hear an impoverished, atypical, or disorganized story and are asked to recall it, they can and do produce a canonical version of it. To determine if this "strategic" manipulation of story structure undergoes developmental changes, two experiments were conducted using second and sixth grade children and college students. In the first experiment, 40 subjects at each grade level listened to "normal" or "scrambled" versions of stories and either recalled them as heard or recalled them as good (organized) stories. Results showed that scrambled stories generally depressed recall and that there was a clear improvement with age/grade in the ability to recognize a scrambled story, with second grade subjects performing especially poorly. The second experiment examined two alternative explanations for the poor performance of the second grade students: (1) younger children's memory for material they have just heard is "fragile," and any attempt to operate on it or transform it is doomed because the effort detracts from the effort to hold on to the memory itself; and (2) sequencing techniques needed to reorder a scrambled story are not well mastered in young children. Eighteen second grade children were trained to sequence the propositions of a random story into a canonical form, with the propositions continuously available for inspection, while a control group of 18 subjects received no special training. Results confirmed that second grade students could reorganize their recall only if some training in sequencing were offered them. (FL)