The nigrostriatal projection was studied with a retrograde tracing method (Evans blue, EB) combined with a technique for dopamine histofluorescence. The study, realized in control rats and in animals with 6-hydroxydopamine-induced lesions of the dopaminergic pathway, yielded the following results. (1) In 3 control rats injected with 0.1 microliter of a 10% solution of EB in the center of the caudate-putamen 1 mm anterior to the globus pallidus, 96% of all substantia nigra neurons retrogradely labelled with the dye contained dopamine fluorescence. The remaining ones (average 350 per brain) were devoid of dopamine fluorescence and predominantly found in the posterior 75% of the substantia nigra. These last cells were confined to the upper-half of the pars reticulata. (2) In a series of 6 animals, the cytotoxic agent 6-hydroxydopamine was injected in various locations in the vicinity of either the substantia nigra or the nigrostriatal tract 12--15 days prior to the injections of 0.1 microliter of EB in the same striatal locations as in the controls. Despite a reduction of up to 85% in the number of dopaminergic cell bodies, the substantia nigra of these rats contained the same average number of EB-labelled neurons devoid of dopamine fluorescence. (3) Eight rats received smaller injections (0.1 microliter) of EB in various striatal sites and in two further cases such injections were placed in the globus pallidus to determine more accurately the anatomical location of the dopamine-negative nigral neurons retrogradely labelled with the dye. Following the striatal injections, these cells were found mostly in the upper-half of the pars reticulata and were arranged in longitudinal oriented clusters whose mediolateral location depended on the striatal injection site. Following the pallidal injections, retrogradely labelled neurons devoid of dopamine fluorescence were found in greater numbers and were located in all areas of the pars reticulata. The possibility of retrograde labelling of some nigrothalamic neurons was not entirely ruled out in these two cases. (4) Finally 6 rats received 0.1 microliter injections of EB in various parts of the parietal cortex. In these cases the substantia nigra did not contain any EB-positive dopamine-negative neurons. These results are interpreted as evidence in support of the existence of a topographically organized non-dopaminergic nigrostriatal projection.