Physical activity is often recommended as a method to improve bone health, yet recent evidence suggests that a threshold may exist at which physical activity becomes deleterious to bone mineralization. Indeed, stress fracture injury is related to high running mileage during military training, and is more prevalent in women. PURPOSE: To assess the relationship between athletic activities (running and high impact sports) and lumbar bone mineral density in physically active, age-matched men and women over a 10-year period. METHODS: 48 male and 59 female West Point cadets participated in this longitudinal study. Dual energy X-ray absorptiometry was used to assess lumbar spine (L2-4) bone mineral density (BMD), bone mineral content (BMC), and bone area biannually during college (1989–1993), and 6 years after graduation (1999). Running hrs/week (RH) and hrs/week of impact sports (SP) were averaged over the college years and related to BMD in 1993 for men (BMD = 1.36 ± 0.15) and women (BMD = 1.30 ± 0.12) separately. RH and SP were also averaged over 1993 to 1999 and related to BMD in 1999 (men BMD = 1.31 ± 0.15 and women BMD = 1.31 ± 0.11). The change in fat mass (FM) and lean mass (LM) from 89–93 and 93–99 were also calculated. Multiple linear regression was used to predict BMD in 1993 and 1999, separately, and the change in BMD from 89–93 and 93–99, using average RH, average SP, ΔFM and ΔLM as independent variables. RESULTS: BMD increased significantly for both men (4.7%) and women (2.8%) during the college years. Six years later women maintained their BMD, whereas male BMD significantly decreased (3.8%). The change in FM and LM was not related to the change in BMD. In men, SP from 89–93 (mean 2.9 hrs/wk, range 0–15.2) was related to BMD in 1993 (R = 0.37, R2 = 0.14, P<0.001), and SP (mean 1.1 hrs/wk, range 0–5.0) from 93–99 was related to BMD in 1999 (R = 0.34, R2 = 0.12, P = 0.02). In women, the change in BMD from 89–93 was negatively related to RH (mean 1.8 hrs/wk, range 0.3–8.8) from 89–93 (R = 0.36, R2 = 0.13, P = 0.005) and was positively related to RH (mean 1.5 hrs/wk, range 0–6.3) from 93–99 (R = 0.47, R2 = 0.22, P<0.001). These relationships were significant for BMC but not for bone area. CONCLUSION: In this study, longitudinal changes in BMD from age 18 to 29 were influenced by impact sport hrs/wk for men and running hrs/wk for women. These changes in BMD appear to be due to mineralization and not to changes in bone area. In college-aged women there was a negative relationship between BMD and running hours, which was not observed in men. This may indicate that gender differences exist in the adaptation of bone to physical activity.