Most pediatric exercise intervention studies that evaluate the effect on skeletal traits include volunteers and follow bone mass for less than 3 years. We present a population-based 6-year controlled exercise intervention study in children with bone structure and incident fractures as endpoints. Fractures were registered in 417 girls and 500 boys in the intervention group (3969 person-years) and 835 girls and 869 boys in the control group (8245 person-years), all aged 6 to 9 years at study start, during the 6-year study period. Children in the intervention group had 40 minutes daily school physical education (PE) and the control group 60 minutes per week. In a subcohort with 78 girls and 111 boys in the intervention group and 52 girls and 54 boys in the control group, bone mineral density (BMD; g/cm 2 ) and bone area (mm 2 ) were measured repeatedly by dual-energy X-ray absorptiometry (DXA). Peripheral quantitative computed tomography (pQCT) measured bone mass and bone structure at follow-up. There were 21.7 low and moderate energyrelated fractures per 1000 person-years in the intervention group and 19.3 fractures in the control group, leading to a rate ratio (RR) of 1.12 (0.85, 1.46). Girls in the intervention group, compared with girls in the control group, had 0.009 g/cm 2 (0.003, 0.015) larger gain annually in spine BMD, 0.07 g (0.014, 0.123) larger gain in femoral neck bone mineral content (BMC), and 4.1 mm 2 (0.5, 7.8) larger gain in femoral neck area, and at follow-up 24.1 g (7.6, 40.6) higher tibial cortical BMC (g) and 23.9 mm 2 (5.27, 42.6) larger tibial crosssectional area. Boys with daily PE had 0.006 g/cm 2 (0.002, 0.010) larger gain annually in spine BMD than control boys but at follow-up no higher pQCT values than boys in the control group. Daily PE for 6 years in at study start 6-to 9-year-olds improves bone mass and bone size in girls and bone mass in boys, without affecting the fracture risk.