1. A tendon tap of the biceps femoris tendon was found to evoke a depression of the soleus and tibialis anterior H reflexes with a duration of 300-400 ms and with an onset at a conditioning-test interval of 20-30 ms. It is suggested that the depression is caused by presynaptic inhibition of the terminals of the Ia afferents mediating the reflexes. 2. This possibility was tested by a method in which the H reflex is facilitated by a monosynaptic Ia volley from the quadriceps muscle. The attenuation of this facilitation when another pathway is stimulated is probably caused by presynaptic inhibition of Ia afferents. It was shown that the biceps femoris tendon tap depressed the size of the femoral nerve-induced facilitation of the soleus and tibialis anterior H reflexes. This suggests that the depression of the reflexes by the tendon tap was indeed caused by presynaptic inhibition. 3. To investigate whether the terminals of descending fibres were similarly susceptible to presynaptic inhibition, the stimulation of the femoral nerve was replaced by magnetic stimulation of the contralateral motor cortex. This stimulation has been shown to evoke a facilitation of the tibialis anterior and soleus H reflexes which (within its initial 0.5-1 ms) is probably caused exclusively by direct monosynaptic projections from the cortex to the motoneurones. In contrast to the facilitation evoked by Ia afferents, the descending facilitation was not influenced by the biceps tendon tap. 4. Similarly, the monosynaptic peak in the post-stimulus time histogram (PSTH) of single voluntarily activated tibialis anterior motor units evoked by stimulation of the common peroneal nerve was depressed by the tendon tap, whereas this was not the case for the presumed monosynaptic peak evoked by brain stimulation. 5. It is suggested that the tendon tap evoked presynaptic inhibition of the terminals of flexor as well as extensor Ia afferents terminating on both soleus and tibialis anterior motoneurones. In contrast, the tap failed to elicit any presynaptic inhibition of the terminals of descending fibres on the motoneurones. We suggest that descending pathways in general are free from the presynaptic control which attenuates peripheral input to motoneurones. Through modulation of presynaptic inhibition the brain may thus selectively hinder the access of peripheral feedback mechanisms to the motoneurones, while still maintaining control of the output from the spinal cord through direct and indirect projections to the motoneurones.