Sixty-one normally achieving and at-risk fifth-grade students (of whom three had learning disabilities), in three classrooms, were taught two 5-week science units via experimental or control conditions in which treatment order and unit of instruction were counterbalanced. In the control condition, students received typical instruction, with teacher lecture and discussion, textbook reading, and worksheet exercise completion. In the experimental condition, students received differentiated curriculum enhancements, in which multitiered activities were undertaken by students in a classwide peer tutoring format. Analysis of gain score data revealed that students scored higher on production tests, but not identification tests, when in the experimental condition. Student and teacher reports indicated a high degree of satisfaction with experimental methods and materials. (Contains 1 figure and 3 tables.)