The benzamide derivative [125I]iodosulpride was used to generate light microscopic autoradiograms on sections of rat brain and spinal cord. Sites specifically labelled by [125I]iodosulpride over a low background correspond to dopamine D-2 receptors as shown by their pharmacology established by densitometric analysis of 11 typical areas from autoradiograms generated in the presence of five dopamine-competing agents. An atlas of D-2 receptors was established using 1 horizontal, 6 sagittal and 30 frontal sections, the latter serially prepared at 0.5-1 mm intervals. Labelled areas were identified by comparison with corresponding, classically stained sections. When their density, rated according to an arbitrary scale, was then compared to that previously reported for dopamine innervation, evaluated from distributional maps of dopamine histofluorescence or tyrosine hydroxylase immunoreactivity, three situations were found. In areas corresponding to cells of origin and established projection fields of the mesostriatal, mesolimbocortical, diencephalospinal and periglomerular systems the density of D-2 receptors generally paralleled that of dopamine innervation. D-2 receptors in substantia nigra (pars compacta or reticulata) and ventral tegmental area were strongly reduced after injections of the neurotoxin 6-hydroxydopamine into the medial forebrain bundle, suggesting their major localization on dendrites and perikarya of dopamine neurons. Most other described dopamine cell group areas also contained D-2 receptors. In contrast many areas without established dopamine innervation contained D-2 receptors, sometimes in high density. This was the case for large areas of the cerebral cortex (layers I-III and V-VI) outside the established projection fields of the mesocortical system, the cerebellum (moleculare layer and dense patches within lobule 9), the hippocampal formation (lacunosum moleculare layer), several septal, thalamic and hypothalamic nuclei, large tectal areas, numerous brainstem areas (including cranial nerve nuclei), etc. This situation might correspond to areas with minor and still undetected dopamine innervation or to a localization of D-2 receptors on cells (or cell parts) not receiving dopamine inputs. Finally several well-established dopaminergic areas did not reveal any D-2 receptor labelling. This was particularly the case in the hypothalamus (areas of origin or termination of the tuberohypophyseal and incertohypothalamic dopamine systems) but also in the hippocampal formation (alveus, fimbria, hilus dentate gyrus), amygdaloid complex (anterior, basolateral, medial nuclei).(ABSTRACT TRUNCATED AT 400 WORDS)