The reliability of outcome measures obtained using the Balance Master and the limits of stability in anterior, posterior, and lateral directions were evaluated in 70 healthy subjects aged 20 to 32 years. Data relating to static sway and the ability to shift the centre of gravity to preset targets were collected on three occasions one week apart. The centre of gravity position and limits of stability were determined over three trials and data converted from a relative reference system to absolute displacements from vertical. Intraclass correlation coefficients revealed fair to poor reliability of static and dynamic sway measures (coefficients < or = 0.55) and excellent reliability of limits of stability measures and the position of the centre of gravity (coefficients > or = 0.75). The variability in outcome measures from tests which do not maximally challenge the postural control system may be a hallmark of normal balance performance. Further, the intersubject variation in resting centre of gravity position and in limits of stability supports the use of absolute performance measures as the interpretive value of data expressed relative to standard norms is limited.