The purpose of this study was to examine the bone mineral densities (BMD) of female junior and senior football (soccer) players with different training regimens and histories, female former football players, and their respective controls. Active junior (age 13-17 years, n = 62) and senior (age 18-28 years, n = 34) players, representing three teams with different levels of performance and training, were compared reciprocally and with matched controls (n = 90). Former players (age 34-84 years, n = 25) who had ended their careers on average 9.7 years previously and their matched controls (n = 57) were also studied. Body composition and total body, lumbar spine and proximal femur BMD were measured with dual-energy X-ray absorptiometry. Former players and their controls were asked in a questionnaire to specify their current level of physical activity. In a control for differences in age, weight and body mass index, football players had significantly greater BMD than controls at all sites measured. This difference appeared to be site-specific, with greater differences in BMD at the proximal femur sites (10.5-11.1%) than at the lumbar spine (4.8%) or for the total body (3.5%). Further, differences were greater for senior than for junior players. However, no BMD differences were found between teams representing different levels of performance and training. Female former football players had retained their proximal femur and total-body BMD advantage over controls. In conclusion, training in female football, which is an impact-loading activity, has a site-specific, positive effect on bone formation that is not increased over a certain level of physical activity. The BMD advantage attained appears to be preserved to some extent after the termination of the athlete's active career, which may have a positive effect on future fracture risk.