Objective: It is generally believed that exercise positively influences bone mineral density (BMD). Athletes have been found to have higher BMD than controls but it has proven difficult to reproduce these findings in the general population. Results from cross-sectional studies on the relationship between exercise and BMD in postmenopausal women have been contradictory. In most studies the age range of subjects has been quite large. Few studies have concentrated on this relationship in elderly women, the largest risk group for osteoporosis and little is known if, and in that case what kind of, exercise has positive effects on BMD in these women. The purpose of this study was to examine the relation of BMD to exercise and current and lifetime occupational activity in 70-year-old Icelandic women. Material and methods: 248 women, all inhabitants in Reykjavik were investigated. BMD in the lumbar spine, femoral neck, total hip and total body was measured with dual energy X-ray absorptiometry (DXA) and the women filled out a questionnaire regarding general health issues, leisure time and occupational activity. Questions included number of leisure walks per week, frequency of other exercises and an attempt was made to estimate the intensity of the activities. Occupational activity was evaluated at ages 20-29 years, 30-44 years, 45-65 years as well as currently, and defined in four grades, from "mostly sedentary" to "hard work including walking". Results: No relationship was found between number of walks and BMD. Significant positive correlation was found between number of other exercise sessions per week and total body BMD (b=0.008, p=0.01), but not total hip (p=0.09), femoral neck (p=0.15) or lumbar spine (p=0.07). Significant negative correlation was found between number of leisure walks and height loss from the age of 25 years (r=-0.211, p=0.001). No significant relationship was found between occupational activity and BMD. Conclusion: Results indicate that leisure time exercise can bring on some bone density benefits for elderly women. Leisure walking alone may not provide high enough stimuli to influence BMD but increasing number of other exercise sessions per week has positive relations to total body and possibly total hip and lumbar spine BMD. A randomized controlled study on the relationship between exercise and BMD in this age group should be conducted.