This paper describes findings from a Follow-Through Project in 2 Napa County, California, schools over a 4-year period. The study explored relationships between program implementation and student engaged rate and achievement in reading and mathematics from 1982 to 1985. Madeline Hunter's Instructional Theory into Practice model was the major training component for teachers in grades 1-4. The study consisted of 4 phases. This paper reports findings from the fourth and final (1984-1985) phase, which consisted of program maintenance by principals and teachers at the schools, with minimal assistance from project staff. Teacher implementation of the model was measured using the Instructional Skills Observation Instrument (ISOI). From 1982 through 1984, there were significant increases in teacher ISOI scores for both reading and mathematics. In 1985, 7 of 11 teachers' reading and mathematics ISOI scores decreased. Student engaged rates in reading and mathematics increased significantly from 1982 to 1984 but dropped significantly in 1985. The longitudinal effects of the program on student achievement were analyzed using 102 students who remained in the program for all 4 years. Students gained significantly from baseline until 1984 (the final year of implementation) and then decreased in 1985, significantly in reading and nonsignificantly in mathematics. Comparisons with matched control schools on standardized achievement tests showed greater gains among control students from 1984 to 1985 in both reading and mathematics. An analysis of achievement and engaged rates for limited-English-speaking students showed that these students gained more than English-speaking students. Teachers' ISOI scores were significantly correlated to students' engaged rates in all but reading in 1985. Inconsistencies were found in correlations between teacher behaviors and student reading achievement in all years of the study. Mathematics achievement was significantly correlated with teacher implementation in 1983 and 1984, but, in 1985, achievement was correlated with only 1 component (guided practice) of the ISOI. Evidence is not strong for a link between implementation of the Hunter model and student achievement. Explanations are offered for the decrease in 1985 teacher implementation and student achievement and engaged rates, and implications for staff development are discussed.