Objective. To assess the feasibility and reliability of ankle plantar flexor stiffness measurements in hemiplegia. Design. Repeated measurements in five consecutive weeks. Background. In hemiplegia, an equinovarus positioning of the foot might be caused by an increased stiffness of the m. triceps surae. Methods. In eight hemiplegic patients the net joint torque of passive muscle stretch was measured as a function of ankle-angle by a dynamometer, at both sides. Ankle-stiffness was characterised and also a biomechanical model of the passive muscle was fitted. Results. In the vast majority of measurements it was possible to obtain measurements that were not distorted by involuntary muscle contraction. These measurements showed for the angle at which a passive plantar moment of 10 N m was reached a standard error of measurement of less than 2.7°. The muscle model showed the increased stiffness as a shortening of the muscle-fibre-length. Conclusion. The feasibility of this method to measure muscle stiffness was fair to good in hemiplegic patients. Provided the abandoning of involuntary muscle activity, the reproducibility warrants application in clinical practice at an individual level. The use of the model relates this changes to a shortened m. soleus and/or m. gastrocnemius. Relevance: Effective clinical decisions for treatment of equinovarus positioning of the foot in the hemiplegic individual, should consider excessive involuntary contractions of the m. triceps surae complex (i.e., spasticity), as well as shortened muscle tissue resulting in high stiffness. Despite the importance of increased stiffness there have been no validated methods of measurement. (C) 2000 Elsevier Science Ltd.