This document examines eighteen studies on the subject of cross-grade tutoring, the studies being classified according to the manner in which the tutoring task is selected--whether based on the needs of the teacher, the tutor, or the tutee. Presentation of the studies follows the following sequence: (1) a description and analysis of experiments run for short periods of time by persons external to the schools; (2) projects run by school personnel ("field studies"); and (3) "programmed tutoring," in which the tutor administers a predesigned program of instruction, and which is usually conducted for the benefit of the tutee. Analyses of the reports concentrates mainly on cognitive learning, though a brief survey of noncognitive outcomes is included. (MJB)