Neural grafting has over the last decade emerged as a possible tool for the substitution of damaged neurons in the central nervous system and for the promotion of symptomatic recovery after brain damage. Transplantation studies in the 6-hydroxydopamine lesion rat model of Parkinson's disease were initiated in the late seventies. The first studies were based on the neuronal replacement paradigm, using developing dopamine brain cells obtained from the substantia nigra region of embryonic cadavers. When implanted into the striatum such grafts were found to reinnervate part of the previously denervated striatum and restore dopamine turnover and release to near-normal levels. In both rats and monkeys the nigral grafts have been shown to normalize some, but not all, Parkinson-like symptoms in the dopamine deficient recipients. Grafting of adrenal medullary tissue was introduced in the early eighties as an alternative to the use of embryonic cadaver tissue. The adrenal medullary grafts have, however, so far shown poor long-term survival in both rats and monkeys, and consistent with this no sustained dopamine release have been observed in the brain of long-term grafted animals. Likewise, no long-lasting effects of adrenal medullary grafts on spontaneous motor or sensori-motor behavior have so far been documented in either the rat or the monkey model. The results so far reported from trials using adrenal medullary grafts in patients with Parkinson's disease appear to conform to the available animal experimental data at least in two important respects: significant long-term graft survival has not been possible to document, and any clear-cut functional effects consistent with sustained graft-induced dopamine release have not been demonstrated. Initial results from ongoing trials using grafts of fetal nigral tissue are presented and discussed.