The aim of this study was to identify, at the ultrastructural level, the neuronal targets of dopamine afferents to the medial prefrontal and the anterior cingulate cortex of the adult rat. Since, in addition to pyramidal neurons, the cortical neuronal population mainly consists of GABAergic nonpyramidal intrinsic neurons, the simultaneous visualization of both dopamine- and GABA-containing neurons should leave the pyramidal neurons as the only unlabelled dopamine postsynaptic target. In this context, we used a double labelling immunocytochemical procedure: a pre-embedding PAP immunostaining to visualize monoclonal conjugated-dopamine (DA) antibody, followed by postembedding immunogold staining with a polyclonal conjugated-GABA antibody. In a single section sampling of 369 DA-immunoreactive (DA-IR) varicosities observed and the GABA-containing elements, 75% of the DA-IR terminals showed no indication of any contact with a GABA neuron. Twenty-five per cent were found in nonsynaptic contiguity with a GABA-immunoreactive neuronal element: axon, dendrite or cell body. When a DA varicosity was in nonsynaptic contiguity with a neuronal perikaryon (5% of cases), this cell was GABA positive. Ten per cent of the DA varicosities were contiguous to a GABA axon, but axoaxonic synapses in either direction were never observed. A symmetrical synapse between a DA varicosity and a GABA-containing dendrite was observed only once. The other 13 DA-IR terminals exhibiting a clear synaptic junction were apposed to nonGABA-containing dendrites, spines and shafts. Triads were observed in which a DA varicosity, forming or not a symmetrical synapse, was apposed to an unlabelled dendrite already receiving a symmetrical junction from another unlabelled axon. These data confirm and extend previous results designating the pyramidal cell dendritic tree as the main synaptic target of DA cortical afferents in rat and primate cerebral cortex. However, a direct effect of dopamine on a subpopulation of intrinsic GABA neurons cannot be excluded.