to explore the relationship between risk of falling at age 90+ and prior physical activity at age 60-70s.population-based cohort study (The 90+ Study).California retirement community.of 1596 cohort members, 1536 had both falls and prior activity data. Mean age = 94 years; 78% female; 99% Caucasian.time spent in active physical activity was self-reported in 1980s; medical history, medication, assistive devices, residence type, and falls (outcome) was collected in 2000s. Activity/fall relationships were assessed using logistic regression.falls were reported by 52% of participants, recurrent falls by 32%, and severe injury by 21% of fallers. In univariate analyses risk of falling at age 90+ was significantly related to medical history (heart disease, TIA/stroke, arthritis, vision disease, depression, dementia), medication use (hypnotics, anti-psychotics, anti-depressants), use of assistive devices (cane, walker, wheelchair), residence type (living with relatives, sheltered living), and source of information (self-report vs informant). Risks of falling and recurrent falls at age 90+ were 35-45% lower in those reporting 30+ minutes/day of active physical activity at age 60-70s compared with no activity. The odds ratio of falling was 0.65 (95% CI = 0.44-0.97) for 30-45 minutes/day and 0.64 (0.44-0.94) for 1+ hour/day adjusting for age, sex, medical history (stroke/TIA, vision disease, depression), use of assistive devices, and source of information.falls are extremely common among the oldest-old and a significant proportion lead to severe injury. This work is the first to show an association between exercise at age 60-70s and lower risk of falling at age 90+.