1. A neuronal relay for input from group II afferents of hindlimb muscle nerves has been found in the previously little explored sacral segments of the cat spinal cord. 2. Electrical stimulation of group II muscle afferents of a number of nerves evoked negative potentials on the surface (cord dorsum potentials) and population postsynaptic potentials (field potentials) within the sacral segments. The largest potentials were evoked by stimulation of the posterior biceps-semitendinosus and triceps surae nerves which evoke much smaller potentials in other segments. Group II afferents of other nerves, notably those which have their main relay within the middle lumbar segments, were much less effective. 3. The sites at which cord dorsum and field potentials evoked by group II muscle afferents were recorded varied in relation to the external topography of the L7-S2 spinal segments but were consistent in their location relative to the pudendal motor nucleus (Onuf's nucleus). Potentials evoked by group II afferents of the posterior biceps and semitendinosus nerves peaked at a level corresponding to the rostral half of Onuf's nucleus and potentials evoked by afferents of the gastrocnemius nerves peaked just rostral to this nucleus. The largest field potentials (of 0.5-1.0 mV) were recorded within the dorsal horn. Field potentials in the intermediate zone were much smaller (< 0.3 mV) and were seen less frequently. 4. Evidence was obtained that the dorsal horn field potentials are to a great extent evoked monosynaptically by the fast conducting fraction of group II muscle afferents: (i) they were evoked at short latencies (2.4-2.7 ms from the stimulus; 1.3-1.7 ms from group I components of afferent volleys and 0.5-0.7 ms from group II components of these volleys), (ii) the conduction times of impulses in the fastest conducting fraction of group II afferents, between the sacral segments (where these impulses were induced by intraspinal stimuli) and the peripheral nerves, were only about 0.5 ms shorter than the latencies of field potentials recorded at the site of intraspinal stimulation and evoked by stimulation of the same peripheral nerves and, (iii) the field potentials followed repetitive stimuli without temporal facilitation. 5. Negative cord dorsum and field potentials were also evoked by small stretches of the semitendinosus and triceps surae muscles. Although they were smaller than potentials evoked by electrical stimulation of sensory fibres and appeared at longer latencies, their presence is consistent with a contribution of muscle spindle afferents to the actions of group II muscle afferents within the sacral segments.(ABSTRACT TRUNCATED AT 400 WORDS)