A critical review of available reports on the epidemiology of fetal alcohol syndrome among American Indians, Alaskan Natives, and Aboriginal peoples of Canada was completed. A search of Medline, the National Institute on Alcohol Abuse and Alcoholism Database, and other relevant data bases was conducted. The reference lists of several publications on fetal alcohol syndrome were reviewed, and four prominent researchers and four government agencies were contacted to identify unpublished articles. This search identified 10 studies, 8 of them cross-sectional. Four of these studies used primary data from the authors' evaluations of children suspected of having fetal alcohol syndrome; the other six used secondary data. The prevalence of fetal alcohol syndrome in the American Indians of the United States and Aboriginal peoples of Canada was consistently high across the 10 studies. These studies have significant restrictions which limit both the confidence in the rates reported and the generalizability of the results. Three studies used data from the province of British Columbia. No study evaluated all children in the study area. Only two studies reviewed death certificates. In only one study were examiners blinded to maternal alcohol use, and no study presented evidence on the sensitivity and specificity of either the screening efforts or diagnostic criteria. Such evidence is especially important in studies of secondary data and in studies that report rates for newborn populations. Studies of the sensitivity and specificity of both the screening and diagnostic criteria for fetal alcohol syndrome would be useful areas for further study. Other study designs, including longitudinal cohort studies, are needed. Additional studies of populations of the American Indians, Alaskan Natives, and Aboriginal peoples of Canada, where low rates of fetal alcohol syndrome are suspected, should be completed. Reviews of death certificates may also bea potentially important source of cases.