A project for studying the prevalence of hereditary ataxias (HA) and familial spastic paraplegias (FSP) in Portugal was set up in 1993. The ascertainment of patients in previous prevalence studies relied mainly on the information of hospital admissions and out-patient contacts with the neurology and other related departments at central hospitals covering the whole region surveyed. Many patients might be overlooked if large populations were studied using this method, since registers at central hospitals are very incomplete and for most part not yet computerized. On the other hand HA and FSP are rare diseases appearing in family clusters, and it would be unreasonable to undertake a sample survey based upon a suitable frame of the Portuguese population. Therefore we decided to carry out a two-phase prevalence survey at district level, involving the collaboration of all physicians working in the district health institutions and the population, in the screening of eligible subjects in phase 1. All subjects screened as positive were examined by a neurologist in phase 2. This method provided a direct estimate of false positives and false negatives were all patients also examined in phase 2, who came to our knowledge using other sources of information. The prevalence of hereditary ataxias and spastic paraplegias in the pilot district was 6.4 per 100,000 inhabitants. The sensitivity of the screening procedure was 81.2% and the predictive value of a positive screening was 25%. Considering the geographically circumscribed district nature of the populations to be studied, the comprehensive sources of case identification used and the high adherence of the health professionals involved, we believe that this method can be widely used, particularly in countries with similar health care services.