Parkinson disease (PD) is characterized by loss of the A9 nigral neurons that provide dopaminergic innervation to the striatum. This discovery led to the successful instigation of dopaminergic drug treatments in the 1960s, although these drugs were soon recognized to lose some of their efficacy and generate their own adverse effects over time. Despite the fact that PD is now known to have extensive non-nigral pathology with a wide range of clinical features, dopaminergic drug therapies are still the mainstay of therapy, and work well for many years. Given the success of pharmacological dopamine replacement, pursuit of cell-based dopamine replacement strategies seemed to be the next logical step, and studies were initiated over 30 years ago to explore the possibility of dopaminergic cell transplantation. In this Review, we outline the history of this therapeutic approach to PD and highlight the lessons that we have learned en route. We discuss how the best clinical outcomes have been obtained with fetal ventral mesencephalic allografts, while acknowledging inconsistencies in the results owing to problems in trial design, patient selection, tissue preparation, and immunotherapy used post-grafting. We conclude by discussing the challenges of bringing the new generation of stem cell-derived dopamine cells to the clinic.