Intrastriatal transplantation of embryonic dopaminergic neurons is currently explored as a restorative cell therapy for Parkinson disease (PD). Clinical results have varied, probably due to differences in transplantation methodology and patient selection. In this review, we assess clinical trials and autopsy findings in grafted PD patients and suggest that a minimum number of surviving dopaminergic neurons is required for a favorable outcome. Restoration of [18F]-fluorodopa uptake in the putamen to about 50% of the normal mean seems necessary for moderate to marked clinical benefit to occur. Some studies indicate that this may require mesencephalic tissue from 3-5 human embryos implanted into each hemisphere. The volume, density and pattern of fiber outgrowth and reinnervation, as well as functional integration and dopamine release. are postulated as additional important factors for an optimal clinical outcome. For neural transplantation to become a feasible therapeutic alternative in PD, graft survival must be increased and the need for multiple donors of human embryonic tissue substantially decreased or alternate sources of donor tissue developed. Donor cells derived from alternative sources should demonstrate features comparable to those associated with successful implantation of human embryonic tissue before clinical trials are considered.