Quantification of joint space width of the ankle could provide information essential to evaluate the effects of potential disease-modifying agents and adverse effects of devices intended to ameliorate osteoarthritis elsewhere in the lower extremity. Current methods require proprietary software or have not been well validated; our purpose was to develop and assess the reliability of a digital joint space width quantification method using public access software. We studied 95 patients, asymptomatic in the ankles and without history of ankle trauma, but with symptomatic medial knee osteoarthritis, participating in an ongoing longitudinal trial. Weightbearing anteroposterior radiographs of the ankle and supine radiographs of the pelvis were assessed, and the narrowest medial and lateral tibiotalar joint space widths and hip joint space widths were measured using Image J software (US NIH, Bethesda, MD). Medial joint space widths were 2.56 +/- A 0.50 and 2.55 +/- A 0.48 mm, and lateral joint space widths were 2.45 +/- A 0.55 and 2.44 +/- A 0.52 mm, for right and left ankle, respectively. Coefficients of variation for repeat measurements by the same observer were 1.13% and 4.5%, and by different observers 7.30% and 7.27%, for medial and lateral joint space widths, respectively. Men had wider joint space widths than women when accounting for height. Joint space width of the ankle correlated with the joint space width of the hip and with height and weight, but not with age.