We have compared the finger hyperextensometer (Jobbins et al., 1978) and the scoring system of Carter and Wilkinson (1964), modified by Beighton et al. (1973), in the quantification of joint laxity in a population selected to demonstrate a wide range of joint movement. Both systems correlated well (P less than 0.001) with a 'global index' of joint laxity derived by adding the arcs of movement recorded at most joints in the body by the method of the American Academy of Orthopaedic Surgeons (1965). The Carter and Wilkinson system produced better correlations and is the method of choice for assessing generalized joint laxity.