Objective: To compare the predictive value of a variety of adiposity measures for the risk of disability. Design/setting: This study used 14-year follow-up of the Melbourne Collaborative Cohort Study (n = 7142). Adiposity measures were collected at baseline and disability measures for 5 self-care activities and mobility were collected at follow-up (2003-2007). Methods: Logistic regression was used to analyze the association between each adiposity measure (body mass index [BMI], waist circumference [WC], hip circumference, waist-to-hip ratio, fat mass, fat free mass, and percentage fat and disability. Area under the receiver operating curve ranking and comparison between nested models were used to determine the best predictor of disability. Results: For men and women, the odds for disability increased with increasing adiposity. In men, BMI was the most predictive adiposity measure for all types of disability. In women, 2 adiposity measures (BMI and WC) predicted overall and mobility disability better than only one measure, with hip circumference the single best predictor for self-care disability. Conclusions: BMI and WC predicted disability well in men and women. Identifying individuals at high risk of future disability through simple measures of adiposity will be essential if we are to adequately cater for our ageing population. ? 2012 Elsevier Inc..