1 The effects of pentobarbitone (PB) and other sedative/hypnotic drugs have been examined in relation to y-aminobutyric acid (GABA) in vitro on the superfused isolated superior cervical ganglion of the rat and in vivo on single units in the brain stem of the anaesthetized rat. 2 PB, and other barbiturates, depolarized the ganglion in a dose-dependent manner (threshold concentration 100-300 pM, cf. GABA depolarization threshold 1 lIM). The depolarization was reduced in the presence of the selective GABA antagonist (+ )-bicuculline methochloride (Bic). Other non-barbiturate sedatives e.g. chlordiazepoxide, amitriptyline, promethazine at concentrations up to 2mM produced no depolarization.3 PB, tested at concentrations up to 80 gM, produced variable effects on the dose-response curve to GABA. On most occasions a slight potentiation occurred in responses to low concentrations of GABA (below 10 JM) coupled with a depression in the responses to concentrations of GABA greater than 10 pM. 4 Superfusion with PB in the presence of Bic reversed the depression in the response to GABA produced by Bic. This reversal phenomenon occurred at concentrations of PB too low to depolarize the ganglion and was dependent not only on the concentration of PB but also on that of Bic. 5 The reversal potency within an homologous series of barbiturates increased with the size of the alkyl substituent (R2) at C5 on the barbiturate ring. The most potent occurred when the substituent contained 5 carbon atoms (pentobarbitone and amylobarbitone); above this, activity decreased. 6 PB reversed the effects of the other GABA antagonists, tetramethylenedisulphotetramine and isopropyl bicyclophosphate and also the non-selective antagonism produced by strychnine. A concomitant reduction by strychnine of responses to the cholinomimetic, carbachol, was not reversed by PB. 7 Non-barbiturate sedative/hypnotics also reversed the GABA antagonism produced by Bic. The benzodiazepines were effective at lower concentrations than PB (chlordiazepoxide threshold concentration 0.5 JIM, cf. PB 5 gM), however, they only produced a partial reversal even at concentrations much higher than the maximally effective concentration of PB. 8 The Bic reversal effect of chlordiazepoxide (and other benzodiazepines) lasted many hours after removal from the superfusion solution. By contrast the effect of PB lasted only 15-30min after its removal.* 9 Chlordiazepoxide (30 gM) applied in the absence of Bic did not affect the response to GABA but did reduce the depression produced by the subsequent application of Bic even though the chlordiazepoxide had been removed 40 min earlier. 10 In the rat brain stem in vivo PB, applied iontophoretically in amounts which neither decreased the spontaneous neuronal firing rate nor affected the response to GABA or glycine, reversed the GABA antagonism induced by iontophoretic application of Bic (in all 23 neurones tested). PB also reversed the antagonism produced by strychnine of responses to glycine although this was less readily observed (5 out...