This paper is on the influences of the classroom context on the course and malleability of aggressive behavior from entrance into first grade through the transition into middle school. Nineteen public elementary schools participated in developmental epidemiologically based preventive trials in first and second grades, one of which was directed at reducing aggressive, disruptive behavior. At the start of first grade, schools and teachers were randomly assigned to intervention or control conditions. Children within each school were assigned sequentially to classrooms from alphabetized lists, followed by checking to insure balanced assignment based on kindergarten behavior. Despite these procedures, by the end of first quarter, classrooms within schools differed markedly in levels of aggressive behavior. Children were followed through sixth grade, where their aggressive behavior was rated by middle school teachers. Strong interactive effects were found on the risk of being highly aggressive in middle school between the level of aggressive behavior in the first grade classrooms and each boy's own level of aggressive, disruptive behavior in first grade. The more aggressive first grade boys who were in higher aggressive first grade classrooms were at markedly increased risk, compared both to the median first grade boys, and compared to aggressive males in lower aggressive first grade classrooms. Boys were already behaving more aggressively than girls in first grade; and no similar classroom aggression effect was found among girls, although girls' own aggressive behavior did place them at increased risk. The preventive intervention effect, already reported elsewhere to reduce aggressive behavior among the more aggressive males, appeared to do so by reducing high levels of classroom aggression. First grade males' own poverty level was associated with higher risk of being more aggressive, disruptive in first grade, and thereby increased their vulnerability to classroom level of aggression. Both boys and girls in schools in poor communities were at increased risk of being highly aggressive in middle school regardless of their levels of aggressive behavior in first grade. These results are discussed in terms of life course/social field theory as applied to the role of contextual influences on the development and etiology of severe aggressive behavior.