1 A common anaesthetic endpoint, prevention of withdrawal from a noxious stimulus, is determined primarily in spinal cord, where glycine is an important inhibitory transmitter. To de®ne pre-and postsynaptic anaesthetic actions at glycinergic snyapses, the e ects of volatile anaesthetic agents on spontaneous and evoked glycinergic currents in spinal cord motor neurons from 6 ± 14-day old rats was investigated. 2 The volatile anaesthetic agents en¯urane, iso¯urane and halothane signi®cantly increased the frequency of glycinergic mIPSCs, en¯urane to 190.4% of control+22.0 (mean+s.e.m., n=7, P50.01), iso¯urane to 199.0%+28.8 (n=7, P50.05) and halothane to 198.2%+19.5 (n=7, P50.01). However without TTX, iso¯urane and halothane had no signi®cant e ect and en¯urane decreased sIPSC frequency to 42.5% of control+12.4 (n=6, P50.01). All the anaesthetics prolonged the decay time constant (t) of both spontaneous and glycine-evoked currents without increasing amplitude. With TTX total charge transfer was increased; without TTX charge transfer was unchanged (iso¯urane and halothane) or decreased (en¯urane). 3 En¯urane-induced mIPSC frequency increases were not signi®cantly a ected by Cd 2+ (50 mM), thapsigargin (1 ± 5 mM), or KB-R7943 (5 mM). KB-R7943 and thapsigargin together abolished the en¯urane-induced increase in mIPSC frequency. 4 There are opposing facilitatory and inhibitory actions of volatile anaesthetics on glycine release dependent on calcium homeostatic mechanisms and sodium channels respectively. Under normal conditions (no TTX) the absolute amount of glycinergic inhibition does not increase. The contribution of glycinergic inhibition to anaesthesia may depend on its duration rather than its absolute magnitude.