Studies have shown that wrist goniometers are prone to measurement errors, particularly due to crosstalk. This study compared two wrist goniometer systems: a commonly used biaxial, single transducer (System A) and a biaxial, two-transducer (System B). Wrist angles, range of movement and crosstalk results were compared. With the wrist in 90 degrees of pronation, eight subjects were placed in 20 different wrist postures between -40 degrees and 40 degrees of flexion/extension and between -10 degrees and 20 degrees of deviation. Relative to System B, System A had larger measurement errors and was more prone to crosstalk. There may be two sources of crosstalk: (1) intrinsic crosstalk associated with the design, application and twisting of the goniometer transducer when on the wrist, and (2) extrinsic crosstalk associated with the anatomy and complex movement of the wrist joint. It appears that the majority of the radial/ulnar crosstalk measured with System A was intrinsic crosstalk due to the twisting of the goniometer transducer. (C) 2001 Elsevier Science Ltd. All rights reserved.