BACKGROUND Prior studies indicate that older women fall more often than men although there is no evidence of gender-based balance differences. Using a force platform, we measured the effects of restricted sensory input and support surface movement to detect gender differences in balance. METHODS Healthy, elderly community dwellers (N = 234, mean age = 76 +/- 5 years, 52% female) were administered the following perturbations on the balance platform: The platform and/or visual surround were fixed or tilted proportionally to the subject's sway with the eyes open or closed, forward or backward horizontal translations, and toes-up and toes-down rotations. RESULTS Gender-based balance differences were not present during quiet standing, or when the support surface or visual input were manipulated separately. Women swayed and lost their balance more than men when the surface was sway-referenced while vision was compromised, but by the third trial their sway control was comparable to the men. Women also initially lost their balance more frequently than men during toes-up and -down rotations, and compared to men continued to lose their balance more often during repeated toes-up rotations. Finally, women developed less angular momentum than men in response to forward platform rotations. CONCLUSIONS Elderly women show impairments of balance when simultaneously deprived of visual and somatosensory inputs or during a backwards destabilization. Since there is little evidence for a CNS source for such gender differences, biomechanical origins (e.g., dorsiflexion strength and range of motion) are a more likely cause. Limited postural control of women under conditions stressing balance may explain their greater frequency of falling.