When fentanyl is given intravenously to anaesthetised rats they become apnoeic due to respiratory depression, and fusimotor neurones are activated with a delay, usually within a min (Gladden & Sahal, 2000; Gladden et al. 2001). Ventilating the rats promptly reduces the fusimotor activity, but if ventilation is stopped fusimotor neurones again become excited with a similar delay. This sequence can be reliably repeated providing that the opiate concentrations are maintained sufficiently to depress ventilation. In some preparations α-motoneurones are also activated, although later than the λ-motoneurones. Opiate-induced excitation of α-motoneurones is supraspinal in origin (Kuschinsky et al. 1977). Supposedly λ-motoneurones could be excited from the same source, but they may perhaps be responding directly to hypoxia. λ-motoneurones in spinalised preparations fire under hypoxic conditions -though note that spinalisation suppresses static λ-activity and activates dynamic λ-motoneurones (Alnaes et al. 1965).

Sprague Dawley rats were anaesthetised with urethane (1.7 g kg^-1 I.P.). In-continuity recordings were taken from a nerve branch to the longissimus caudae muscle, and the S1 dorsal and ventral roots. Multi-unit records were rectified and averaged. Individual λ-and afferent spikes were identified by their shapes and appropriate latencies between the nerve and spinal roots. Following an initial dose of fentanyl (50 µg kg^-1 I.V.) and before spinalisation, responses were gauged during 4 periods of 45 s of apnoea with at least 5 min recovery between -two periods before and one each after a laminectomy at Th2-3 and a hemisection of the cord. The earliest increases in spikes were detected in records from the ventral roots. Any spontaneous ventilation was suppressed with further doses of fentanyl. The animals were sacrificed after the experiments with an overdose of anaesthetic.

Before spinalisation the mean delay between the last respiration before the apnoea and the start of the increased output from the ventral roots was 29 s (S.E.M. 1.5; n = 27). After spinalisation the delay before any response was much longer (mean 70s ± 4.1; n = 5). In each individual experiment the delay after spinalisation was at least twice the longest delay before spinalisation, and in one case there was no response. Thus the excitation of λ-motoneurones with a short delay must be supraspinal in origin. This excitation survived contralateral hemisection (n = 5), but ipsilateral hemisection abolished it in one case and much attenuated it in another, suggesting that the descending pathway is mainly ipsilateral.