We have raised antisera against the GABA analogues gamma-vinyl GABA, diaminobutyric acid and gabaculine. These analogues are thought to be substrates for high-affinity GABA transporters. Retinae were exposed to micromolar concentrations of these analogues in the presence or absence of uptake inhibitors and then fixed and processed for immunocytochemistry at the light and electron microscopic levels. Immunolabelling for gamma-vinyl GABA revealed specific labelling of GABAergic amacrine cells and displaced amacrine cells in retinae of rabbits, cats, chickens, fish and a monkey. GABA-containing horizontal cells of cat and monkey retinae failed to exhibit labelling for gamma-vinyl GABA, suggesting that they lacked an uptake system for this molecule. In light-adapted fish, gamma-vinyl GABA was readily detected in H1 horizontal cells; similar labelling was also observed in light-adapted chicken retinae. The pattern of labelling in the fish and chicken retinae was modified by dark adaptation, when labelling was greatly reduced in the horizontal cells, indicating the activity dependence of GABA (analogue) transport. Intraperitoneal injection of gamma-vinyl GABA into rats resulted in its transport across the blood-brain barrier and subsequent uptake into populations of GABAergic neurons. The other analogues investigated in this study exhibited different patterns of transport; gabaculine was taken up into glial cells, whilst diaminobutyric acid was taken up into neurons, glial cells and retinal pigment epithelia. Thus, these analogues are probably substrates for different GABA transporters. We conclude that immunocytochemical detection of the high-affinity uptake of gamma-vinyl GABA permits the identification of GABAergic neurons which are actively transporting GABA, and suggest that this novel methodology will be a useful tool in rapidly assessing the recent activity of GABAergic neurons at the cellular level.