Routine study of three-dimensional (3D) tarsal kinematics is hampered by errors due to the displacement of skin surface-tracking markers relative to the underlying bones. Reliable kinematics can be obtained with bone-fixed markers, but an accurate, non-invasive method would have more applications. Simultaneous kinematic data from skin-based and bone-fixed markers attached to the tibia and third metatarsus were collected from three trotting subjects. The motion of the skin-based markers was extracted relative to the underlying bone motion tracked using the bone-fixed markers. The 3D skin displacement patterns for the skin-based markers were parameterized using a truncated Fourier series model. These displacements were expressed in terms of the local coordinate system for each bone. Skin displacement artefacts were observed in all three axes of each bone segment, with the largest displacements occurring at the proximal tibia. The mean skin displacement amplitudes in the tibia were 6.7%, 3.2% and 10.5% of segment length, and for the third metatarsus were 2.6%, 1.4% and 3.8% of segment length, for the craniocaudal, mediolateral and longitudinal segment axes, respectively. Skin displacement patterns could be expressed concisely using the Fourier series model. Displacements were also consistent between subjects, which should allow them to be used as a basis for developing a correction procedure for 3D tarsal joint kinematics.