A primary neuronal culture was prepared from the ventral mesencephalon, centered on the A8, A9 and A10 dopaminergic nuclei of the embryonic day 14 rat, and studied from 12 h to 28 days. At 12 h after plating, and before cell death ensued, 95% of the cells stained positive for neuron specific enolase; 20% for tyrosine hydroxylase; 5% for vimentin and < 0.1% for glial fibrillary acidic protein. In the presence of the mitotic inhibitor cytosine arabinoside (2.0 microM), neuronal growth and survival were surprisingly normal up to the ninth day in culture, but deteriorated rapidly thereafter. In the absence of a mitotic inhibitor, and in the presence of proliferating but non-confluent glia, the tyrosine hydroxylase positive neurons that survived to the 10th day, had retracted neurites and a rounded soma, suggesting an inhibition of cell development. Those tyrosine hydroxylase positive neurons that survived this adverse phase of development tended to produce elaborate neuritic profiles after the 11th day, coincident with confluence of the astrocyte monolayer at the 12th day. By the 21st day in culture, and persisting up to the 28th day, 60% (61 +/- 10, n = 20) of the surviving neurons stained positive for tyrosine hydroxylase. When plated on an established, ventral mesencephalic monolayer of astrocytes, at the seventh day in culture, neuritic growth and branching of the tyrosine hydroxylase positive neurons were greater, compared with similar neurons grown on poly-D-lysine, and the signs of arrested development (retraction of neurites and rounded soma) seen at the 10th day after plating on poly-D-lysine, were not observed. We conclude that in the primary culture studied, and under the experimental conditions used, the survival of dopaminergic neurons was independent of glia during the first nine days, and critically dependent on glia thereafter. The resurgence of growth of dopaminergic neurons after 10 days in vitro, and their subsequent selective survival in culture, suggest that confluent type-1 astrocytes produce factors that act selectively on the dopaminergic neuronal phenotype. The successful identification of these dopaminergic-specific, neurotrophic factors could lead to an increased understanding of the etiology of Parkinson's disease, and suggest new directions for therapeutic intervention.