There have been a number of studies looking at the prevalence of Parkinson's disease (PD) in different racial and geographical populations. Some of the earliest studies suggested a difference in the prevalence of PD in African Americans as compared with Caucasians. As such a difference would have important implications for healthcare and research into the etiology of PD, we undertook a review of published studies to determine whether evidence suggested that such a difference exists. We reviewed 20 studies that looked at incidence, prevalence, and percentages of neurology patients with PD and Parkinsonism in Africa and in African-American populations. Two of these were door-to-door studies that relied on questionnaires for initial ascertainment, another was performed by review of outpatient records of a large health maintenance organization, while the remainder were based on hospital admissions, diagnosis in the community, or death certificate reports. In the aggregate, these studies suggest PD may be less frequent among Africans and African Americans than among Caucasians, although the most well-designed study showed only a statistically insignificant reduction in the prevalence of PD among African Americans. Although an apparently lower disease frequency among people of African origin may have a basis in the pathobiology of the disease, nearly all of these studies were vulnerable to a variety of ascertainment biases, and many lacked stringent application of diagnostic criteria applied by specialists trained in movement disorders. We conclude that a difference in the prevalence of PD and Parkinsonism between black and other populations is unproven and will require additional well-designed studies to determine if previously reported ethnic differences in disease prevalence are real.