Relationships of academic achievement (government tests) with physical fitness (multistage run), physical activity (pedometers) and percent body fat (dual emission X-ray absorptiometry) were examined at both the aggregate school level and the individual child level using data collected from 757 children in 29 elementary schools. Statistical adjustments included gender, grade and socioeconomic status. Between-school relationships of the academic scores with fitness and physical activity were strong and positive, with some evidence of (negative) relationships with percent body fat. The between-child relationships were weaker, and nonexistent with percent body fat. Stronger between-school than between-child relationships favor the argument that variation in school cultures, characterized by concurrent attention to fitness and academic achievement, might play a more dominant role in explaining these relationships than any direct effect of fitness on academic achievement.