The major goal of this investigation was to collect statistically-based anthropometry describing the kinematics of the human hand and to model this anthropometry as a function of external hand measurements, so that it may be predicted noninvasively. Joint centres were anatomically estimated as the centre of curvature of the head of the bone proximal to the given joint. Joint centres determined using Reuleaux's method for PIP and DIP were within 1.4 mm of this anatomical estimate. Models using bone length as the independent variable explain more than 97% of the variability in the anatomically estimated joint centre position along the mid-line of the bone. Models for estimating the lengths of the kinematic segments using external hand length as the independent variable account for between 49 and 99% of the variability in segment length. Models for estimating the axial location of the finger MCP and thumb CMC joints with respect to the distal wrist crease using external hand length as the independent variable account for between 82 and 96% of the variability in these locations. Models for estimating the radio-ulnar location of the finger MCP and thumb CMC joints with respect to the long axis of the third metacarpal using external hand breadth as the independent variable account for between 30 and 74% of the variability in these locations.