We measured the impact of diet, anthropometry, physical activity and lifestyle variables on rates of hip bone mineral density (BMD) loss in 470 white men and 474 white women aged 67-79 years at recruitment dwelling in the community. The subjects were recruited from a prospective population-based diet and cancer study (EPIC-Norfolk) in Eastern England. Dietary intake was measured at baseline using 7-day food diaries and used to calculate intakes of some 31 nutrients and 22 food groups. Standardised questionnaires were used to collect data on anthropometry, physical activity and lifestyle variables. BMD loss (percent per annum; % p.a.) was measured using dual-energy X-ray absorptiometry performed on two occasions an average of 3 years apart (range 2-5 years). The mean rate of BMD change at the total hip region was -0.17% p.a. (SD 1.3% p.a.) in men and -0.41% p.a. (SD 1.2% p.a.) in women. In both men and women, weight gain protected against (and weight loss promoted) BMD loss ( P<0.0001). Markers of current physical activity were protective. In men, an increase of 1 l/s in FEV(1) was associated with an increase in BMD at an average rate of 0.25% p.a. ( P=0.013). In women, for every ten trips made per day climbing a flight of stairs, BMD increased at a rate of 0.22% p.a. ( P=0.005) and additionally a 10% increase in activities of daily living score was associated with BMD increasing at a rate of 0.12% p.a. ( P=0.011) in women. Nutritional variation appeared to have less impact on BMD loss. In men there was no evidence of an effect of any of the nutrients evaluated. However, in women, low intake of vitamin C was associated with faster rate of BMD loss. Women in the lowest tertile (7-57 mg/day) of vitamin C intake lost BMD at an average rate of -0.65% p.a., which was significantly faster compared to loss rates in the middle (58-98 mg/day) and upper (99-363 mg/day) tertiles of intake, which were -0.31% p.a. and -0.30% p.a., respectively ( P=0.016). There was no effect of fruits and vegetables, combined or separately, on rate of BMD loss. The results confirm that weight maintenance (or gain) and commonly practiced forms of physical activity appear to protect against BMD loss in this age group. Measures such as ensuring good general nutrition to guard against weight loss in the non-overweight elderly and maintenance of physical fitness could be valuable in protecting against BMD loss. The protective effect of vitamin C in women needs to be further investigated in other prospective cohort or intervention studies.