This study investigated the effect on learning of the interspersing of questions with sections of written discourse. A 5,200-word passage was divided into seven sections, from each of which several completion questions were derived. A pair of questions was inserted before, after, or both before and after the section. These questions also formed an "Old Questions" criterion test; other questions were used to form a "New Questions" criterion test. Forty subjects volunteered from each of two halves of a Spanish-speaking sophomore class (divided above and below the median, according to grade point average). Subjects from each ability level were randomly assigned to one of three conditions of position of inserted questions, or to a control condition with no inserted questions. Results indicated that high-ability students performed better than did low-ability students and that students performed better on the Old Questions test than on the New Questions test. For the Old Questions test, all inserted-questions groups performed significantly better than did the control group, but, for the New Questions test, no inserted-questions group performed better than did the control group. (Author/AA)