An epidemiological study of multiple sclerosis carried out in Gijón (Principado de Asturias). It had the methodological peculiarity of having been performed in a reduced population of 81,462 individuals selected from the primary health care centers from which they depended. This permitted a better case finding and both a prospective and retrospective investigation. In the prospective period, the incidence found was 0.49/100,000/year, while it was 1.59 in the retrospective period. The reasons for this difference are discussed. The prevalence was 23.32/100,000. We attribute the fact that this rate is higher than that previously reported for Spain to methodological and perhaps geographical reasons.