This letter deals with the observed development of (false) twist in fibres subjected to a biaxial rotation-fatigue test. A simple theory in terms of forces and moments is presented from which both the torque and the rate (amount) of twist can be correlated with the viscoelastic properties of the fibre.