Twenty-four teachers, from a school system not using the new elementary science programs, taught representative new units to 544 students, with variations in their inservice training and the amount of classroom science equipment provided. Classrooms at three grade levels (1, 3, and 6) were used twice under opposing conditions. For each unit, the teachers selected for training received a teacher's guide and specific training for the unit. The teachers not selected for training received only a teacher's guide. Some classrooms received sufficient laboratory instructional aids to allow students to work in small groups of two to four students; others received minimal equipment so that their lessons included demonstrations rather than maximal interaction of children with materials. Pupils were given individual oral tests to assess subject matter achievement. Additional equipment resulted in significantly greater achievement only at the first grade level, but all pupils expressed a preference for additional aids. Specific teacher training for the unit did not produce significant differences in pupil achievement. Additional data regarding the effects of sex, I.Q., and family occupation on pupil achievement and pupil preferences are reported. (EB)