The light-evoked release of [3H]acetylcholine (ACh) from the rabbit retina in vivo was measured and taken as an index of cholinergic amacrine cell activity. The light-evoked release of [3H]ACh was reduced by locally applied gamma-aminobutyric acid (GABA), muscimol and 3-aminopropanesulphonic acid (3-APS). The concentrations of these drugs which reduced the light-evoked release of [3H]ACh by 50% (EC50) were 900, 0.3 and 5 microM respectively. In contrast, (-)-baclofen (5 mM), but not (+)-baclofen, significantly increased the light-evoked release of [3H]ACh. The GABA antagonist, bicuculline increased the resting release of [3H]ACh but abolished the inhibitory action of muscimol on the light-evoked release of [3H]ACh. Glycine and taurine also reduced the light-evoked release of [3H]ACh from the retina, their EC50 values being 1.5 and 0.3 mM respectively. This action was blocked by strychnine, but not by bicuculline. In contrast to the GABA antagonist, strychnine did not affect the spontaneous resting release of [3H]ACh. Retinal [3H]ACh release was not affected by dopamine, 5-hydroxytryptamine (5-HT) morphine, substance P, somatostatin, cholecystokinin sulphate, thyrotropin releasing hormone, luteinizing hormone releasing hormone or angiotensin. Electroretinographic changes produced by amino acids and GABA agonists involved mainly the b-wave and were not correlated with their effects on ACh release. Thus, GABA increased the b-wave amplitude, 3-APS had no effect, whilst muscimol, taurine and glycine either had no effect, or reduced the b-wave amplitude. No obvious changes in the e.r.g. were produced by baclofen, dopamine, 5-HT, morphine or any of the peptides studied with the exception of somatostatin, which reduced the amplitude of the b-wave. It is concluded that cholinergic amacrine cell activity in the rabbit retina may be affected by inputs from other amacrines using GABA or glycine (taurine) as their transmitters, but probably not by inputs from peptidergic or dopaminergic amacrine cells. Our experiments do not provide evidence on the sites of action of GABA, glycine or taurine but the action of bicuculline on the resting release of ACh implies that the activity of the cholinergic amacrine cells is affected by a tonically active GABAergic input.