The history of the nigrostriatal dopamine system may provide a prime example of the two faces of scientific development. First, a given concept is replaced by another simply as a result of methodologies being improved, and second, successive technical improvements make seemingly settled controversies even more complicated and disputable. The nigrostriatal pathway, which had been unrecognizable with Nauta's silver impregnation method, became apparent by use of the more sensitive silver impregnation method of Fink-Heimer. The sensitivity of the latter method, however, was still insufficient to reveal the whole extent of another ascending dopamine system, the mesocortical dopamine system, until its existence was established through the application of glyoxylic acid fluorescent histochemistry. Electron microscopic analysis of nigrostriatal dopamine synapses in properly fixed tissue was initiated by the demonstration of dark type terminal degeneration, which was induced by either electrolytic lesions or chemical destruction with a specific toxin (6-hydroxydopamine) of the substantia nigra and medial forebrain bundle. The degenerating terminal boutons, thus produced, invariably formed postsynaptic membrane specializations of asymmetric type. However, the asymmetric nature of the synaptic morphology, although later confirmed by the combined study of chemical lesions and autoradiographic anterograde tracing, was seriously challenged with the introduction of electron microscopic immunohistochemistry. The latter method has consistently revealed that symmetric en passant synapses or axonal varicosities with no synaptic membrane specializations are the only tissue compartments immunoreactive to antibodies against dopamine and its synthetic enzyme tyrosine hydroxylase. In view of the fact that more than 95% of the nigrostriatal projection neurons are dopaminergic, it is difficult to satisfactorily interpret all the available and seemingly paradoxical fine structural data. In this context, a novel concept has emerged in the process of eliminating all the possible alternative interpretations. The concept is that single nigrostriatal neurons form two chemically distinct types of synapses, one dopaminergic symmetric en passant bouton and another non-dopaminergic (still chemically unclassified) asymmetric terminal bouton. If the concept is a valid one, it contradicts Dale's long standing principle, as defined by Eccles: at all the axonal branches of a neuron there is liberation of the same transmitter substance or substances. Furthermore, a certain population of substantia nigra pars reticulata neurons has recently been recognized to be immunoreactive to both dopamine synthetic tyrosine hydroxylase and GABA synthetic glutamate decarboxylase. These single neurons send projections to both the striatum and superior colliculus by way of axon collaterals.(ABSTRACT TRUNCATED AT 400 WORDS)