The jump and kick of the grasshopper are behaviours which are potentially critical for the survival of the animal, and whose maximal performance depends upon optimizing the rate and level of tension development in the extensor tibiae muscle of the hind legs. In experimental conditions extensor tension control can be reduced to a single motoneuron, the fast extensor tibiae (FETi). The axon of FETi can be cut using dye-mediated laser photoaxotomy without damaging the central or peripheral portions of that neuron or any other neuron innervating the leg. The axotomy can be functionally reversed (i.e. the cut axon repaired) by an electronic axonal bypass which detects FETi spikes on the proximal side of the cut and stimulates the axon on the distal side of the cut. In this way motor spikes can either be allowed to reach the muscle or prevented from doing so (by switching the bypass on or off), and the motor programmes produced with and without extensor tension can be compared. The jump and kick are normally produced by a three-stage motor programme: (i) initial flexion brings the tibia into the fully flexed position; (ii) coactivation of extensor and flexor muscles allows the extensor muscle to develop maximal tension almost isometrically, while the simultaneous contraction of the flexor muscle holds the tibia flexed; (iii) sudden trigger inhibition of the flexor system (motoneurons and muscle) releases the tibia and allows the behaviour to be expressed. The grasshopper can produce fictive kicks with motor programmes which show each of these three major structural features of a normal kick, but without any extensor tension whatsoever. There is no significant difference in the frequency of FETi spikes, the duration of coactivation or the maximum depolarization of the flexor motoneurons between fictive and quasi-normal (i.e. reversed axotomy) kicks. The trigger inhibition of flexor motoneurons is shallower in fictive than in quasi-normal kicks. The significance of this is discussed in relation to the activity of the interneuron M, which is known to mediate trigger inhibition onto FITi motoneurons. There are two main conclusions from this study. First, the CNS does not need feedback from ETi muscle tension in order to produce the three-stage motor programme of the kick (and, by implication, the jump). Second, the CNS does not adjust the frequency or duration of FETi activity in response to unexpected changes in ETi tension. ETi tension appears to be under open-loop control in the kick motor programme.