SUMMARY1. The EMG sequence activated before the initiation of a number of fast forwardoriented voluntary movements was analysed quantitatively in normal subjects.2. The sequence consisted of an initial inhibitory component directed to the soleus motor nucleus, followed by a second excitatory one directed to the tibialis anterior (TA).3. The spectrum of functional utilization included motor tasks in which the prime movers are leg and thigh muscles (initiation of gait, rising on tip-toes), thigh and trunk muscles (fast-forward bending of the trunk, standing up) and upper-limb muscles (forward throw or catch).4. In a same motor task and across the different motor tasks, performed at various speeds, the latency of soleus inhibition and TA activation with respect to the onset of movement co-varied according to a linear function, indicating a close temporal correlation between the two components.5. In all the movements investigated, the earliest mechanical effect was a backward displacement of the centre of foot pressure in the sagittal plane.6. Soleus inhibition alone and TA burst alone were each able to produce a backward displacement of the centre of foot pressure, but the effect was significantly slower after soleus inhibition.7. The spatio-temporal parameters of the sequence were modulated according to the pre-existing postural conditions. For the gait initiation protocol, increasing initial forward leaning led to a decrease in the amplitude of soleus inhibition and the TA burst, and to a change in their relative time delays. Modulation was different on the two sides. We could define a postural boundary as the degree of forward leaning beyond which the full sequence is no longer called into action.8. The spatio-temporal parameters of the sequence were pre-set according to the requirements of the forthcoming movement. In the gait initiation protocol, the amplitude and synchronization of the TA burst were directly correlated with velocity of movement, while the relative delay between soleus inhibition and TA activation was inversely correlated. Modulation on the two sides differed. We could define a velocity boundary as the velocity of movement below which the full sequence is no longer called into action.9. We suggest that the EMG sequence described can be considered a motor MS 8303 P. CRENNA AND C. FRIGO programme that, through direct action on the position of the centre of foot pressure (the variable primarily controlled), will precisely adjust the configuration of forces external to the body, allowing the contraction of the prime mover(s) to interact appropriately with them for the production of a specific, forward-oriented movement.