Although recent animal and human experiments suggest that tissue implantation can ameliorate parkinsonism, there is controversy about what mechanism underlies recovery. Secretion of dopamine from the graft seems unlikely to be the sole restorative factor. Regenerative sprouting by the host brain may also underlie behavioral recovery. Fetal amnion and term amnion, which were shown to produce and secrete a factor that supports the outgrowth of neurite processes in vitro, were implanted in hemiparkinsonian monkeys. Fetal amnion implants induced sprouting of dopaminergic fibers from the host brain and behavioral improvement, despite failure of the grafts to survive. Animals implanted with term amnion also had some sprouted dopaminergic fibers and behavioral improvement, but these were limited and were similar to the recovery, in prior experiments using the same primate model of parkinsonism, of animals that received surgical cavitation only. Recovery after central nervous system grafting with fetal amnion, a fetal accessory tissue, does not require secretion of a deficient neurotransmitter(s) from the graft and occurs despite the failure of graft survival. Recovery after cerebral implantation of fetal tissues appears to depend more on the regenerative and recuperative processes of the host brain than on graft replacement of deficient neurotransmitters or development of functional synaptic connections between the graft and the host brain.