Respiratory interneurones in the C4-C6 segments of the spinal cord have only recently been described; until now their projections and connections were not known. We investigated expiratory interneurones in the C5 spinal segment, using antidromic mapping to trace their projections and spike-triggered averaging to test their synaptic connections with phrenic motoneurones. 2. A total of seventy expiratory interneurones were recorded in nineteen cats anaesthetized with pentobarbitone, paralysed and ventilated. The interneurones were found scattered dorsomedial to the phrenic motor nucleus, with discharge patterns of a constant (66%), augmenting (24%) or decrementing (10%) type. 3. Interneurone axons were found in the ipsilateral ventrolateral funiculus using antidromic activation at thresholds < 20 microA. The axons of eighteen of thirty-three interneurones tested (55%) were found to extend to the rostral part of the C6 segment, seventeen of thirty-three (52%) to the caudal part of the C6 segment and ten of nineteen (53%) to the rostral part of the C7 segment. 4. Axon collaterals for thirteen of thirty-three interneurones (39%) were found in the ipsilateral half of the C6 segment, with their endings near the phrenic motor nucleus. In three cases two collaterals were found. None of the interneurones had projections in the contralateral halves of the C5 or C6 segments. 5. In a separate group of thirty-four expiratory interneurones, antidromic mapping was used to find an axon collateral in the C6 segment prior to spike-triggered averaging. Eleven of these interneurones had collaterals (32%) and were subsequently tested for synaptic connections to thirty-two phrenic motoneurones. In three separate instances (9%), inhibitory postsynaptic potentials were observed. Amplitudes, fall times and half-amplitude widths of the inhibitory postsynaptic potentials were 6.7, 10.4 and 10.6 microV; 0.3, 0.5 and 0.7 ms and 0.6, 1.6 and 3.3 ms respectively. 6. We conclude: (i) there is a population of expiratory interneurones in the C5 segment, located predominantly dorsomedial to the phrenic motor nucleus; (ii) at least one-half of these interneurones have ipsilateral intersegmental projections to the C6 segment and (iii) although synaptic connections from expiratory interneurones in the C5 segment to phrenic motoneurones in the C6 segment may be rare, the observed inhibitory postsynaptic potentials had fall times and latencies commensurate with monosynaptic connections.
