This study investigated the effect of various dimensions of classroom structure on children's reading skills. Two classrooms each from grades one to three were selected by district administrators as outstanding examples of different types of classroom structure. A sample of 16 children was drawn from each class. Each classroom was observed on the same two days of the week, for three consecutive weeks, by trained observers using instruments that focused separately on teacher and student behaviors. The Wide Range Achievement Test was administered in the fall and again in late spring. Among the findings were that there were significant relationships between growth in reading and classroom openness, student involvement, and student responsibility for learning. Negative correlations were found between growth in reading and both the amount of teachers' controlling behaviors and positive emotional tone. (AA)