Although previous longitudinal studies suggest bone loss ceases at the spine and radius in women over age 65, few data are available to determine whether femoral bone loss continues in this age group. Because a low femoral bone mass is associated with an increased risk of hip fracture and current therapies for fracture prevention focus on halting bone loss, it is imperative to determine whether femoral bone loss occurs in those who sustain 90% of hip fractures: women over age 65. To determine the annual rate of femoral bone loss, the association of bone loss and age, and the relationship between femoral and spinal bone loss in this age group, we prospectively followed femoral and spinal bone mineral density (BMD) in 85 healthy, community-dwelling, ambulatory women over age 65 (mean 77 years, range 66-93 years). Measurements of femoral and spinal BMD were assessed twice over 1 year using dual-energy x-ray absorptiometry. Cross-sectional analysis of site-specific baseline BMD suggested a significant bone loss in the femoral neck (-0.76% per year, p < 0.01), total hip (-0.70% per year, p < 0.01), trochanter (-0.71% per year, p < 0.05), intertrochanter (-0.88% per year, p < 0.01), and Ward's triangle (-0.86% per year, p < 0.01) but no significant change at the spine. Longitudinal analysis (mean + SD) revealed a similar loss of bone at the total hip (-0.95 + 2.88% per year, p < 0.01), trochanter (-0.74 + 3.62% per year, p < 0.01), and intertrochanter (-1.14 + 3.54% per year, p < 0.01), with a significant gain in lumbar BMD (0.94 + 3.34% per year, p < 0.05). The annual percentage change in BMD at the total hip was negatively correlated with age (p < 0.05). These data demonstrate that femoral bone loss is present in women over age 65 and that measurements of spinal BMD may be a misleading reflection of femoral skeletal integrity in this age group. Furthermore, these data provide the scientific rationale to develop and implement femoral bone loss prevention strategies in this important target group.