Each leg of a standing stick insect acts as a height controller. The leg contains several joints. Most of these joints are known to be controlled by feedback loops which are the basis of resistance reflexes (review Bässler 1983). This leads to the question of whether the resistance reflex of the whole leg can be understood as a simple, vectorial sum of the individual reflexes provided by the different joints, or whether additional properties emerge by simultaneous stimulation of several joints. Force measurements were performed while passively moving the middle leg tarsus of a fixed stick insect (Carausius morosus) stepwise to different positions. From the dynamic and static forces the torques developed by each joint were calculated. They were compared with the torques developed when only a single joint was moved by the same amount. The comparison shows that for a large range of positions there are no differences between both situations. Differences occur in two cases. First, the muscle system controlling the coxa-trochanter joint seems to be more strongly excited when the entire leg is moved than when only the one joint is moved. This change increases the linearity of the whole system for small deviations from the zero position. Second, the torque developed by the extensor tibiae system for negative steps (corresponding to increased body height), and the levator of coxa and trochanter for positive steps, decreases rather than increases when the whole leg is moved to extreme positions. This contributes to a decrease in the slope of the force-height characteristic and thus to a more non-linear behaviour of the whole system for the extreme positions. It is well known that the amplification factors of resistance reflexes in the leg show a large variation (Bässler 1972a; Kittmann 1991). Our results indicate that any change of the amplification factor influences the reflexes in all leg joints in the same way.