A new measure of balance, functional reach, has been recently developed. Functional reach is the maximal distance one can reach forward beyond arm's length while maintaining a fixed base of support in the standing position. Reliability, criterion, and concurrent construct validity of functional reach have been established. The purpose of this study was to assess the predictive validity of functional reach in identifying elderly subjects at risk for recurrent falls. Two hundred and seventeen elderly, community-dwelling male veterans (aged 70-104) underwent baseline screening and were followed for 6 months to monitor falls. Subjects with two or more falls during the 6-month follow-up were classified as recurrent fallers. Logistic regression reveals that if individuals were unable to reach, the adjusted odds ratio (OR) of having two falls was 8.07 (2.8-23.71); if their reach was less than or equal to 6 inches the OR was 4.02 (1.84-8.77); and if reach was greater than 6 inches but less than 10 inches the OR was 2.00 (1.35-2.98). The association between functional reach and recurrent falls was not confounded by age, depression, or cognition. We conclude that functional reach is a simple and easy-to-use clinical measure that has predictive validity in identifying recurrent falls.