When exposed to exogenous serotonin, specific neurones in retinas from different species (guinea pig, rabbit, frog, pigeon and goldfish) take the amine up; these same neurones do not reveal endogenous serotonin following immunocytochemistry for the localization of the amine. The use of autoradiography and immunocytochemistry to localize the uptake of radioactive or unlabelled amine gives identical results, with the exception of the guinea pig retina, where a subset of horizontal-like cells specifically takes up a metabolite of [3H]serotonin (presumably produced in the extracellular space of the retina). This uptake is blocked specifically by N-acetyl-serotonin. After an intraocular injection of forskolin into the rabbit eye the serotonin content of the tissue is elevated and this is reflected in the demonstration of serotonin immunoreactivity associated with a subpopulation of amacrine cells. These cells correspond in size and position to the "indoleamine-accumulating cells" of the retina. Animals (goldfish, frog, pigeon) dark-adapted over a certain period manifest an elevated level of serotonin associated with their retinas compared with that in light-adapted animals. However, immunofluorescence histochemistry used to localize serotonin showed that the same neurones in the different retinas contain the amine, although the "fluorescence" was elevated in the dark-adapted retinas. In the light of the present data it is clear that what constitutes a "serotonergic" neurone cannot be restricted to the immunohistochemical demonstration of the amine. The evidence is persuasive for "serotonergic neurones" having the property to take up serotonin specifically by a high affinity transport process. What warrants serious consideration is whether this property is a characteristic of "serotonergic neurones" rather than the endogenous localization of serotonin without pharmacological pretreatment.