The only measures of morbidity for Guillain-Barré syndrome in well-defined populations in the continental United States come from Olmsted County, Minnesota, and San Joaquin County, California. The estimated incidence rates are between 1 and 2 cases per 100,000 population per year. Despite the hazards of generalizing from such small samples, these are the only estimates available. Large case series have suggested that the syndrome is more common in males, that it may occur at any age, and that it is frequently associated with some antecedent infection. Only two case-control comparisons have been carried out which reveal an association between Guillain-Barré syndrome and an infection 1 month prior to onset of neurological symptoms. A recent survey of four states in the United States revealed that the risk of Guillain-Barré syndrome among individuals receiving immunization against influenza A/New Jersey is 7.3 times the risk among the nonvaccinated. Aside from these investigations and one study linking Guillian-Barré syndrome and high antibody titers to cytomegalovirus, adequate epidemiologic studies to test suggested etiologic hypotheses have not yet been carried out. The rarity of the condition and the variations in its clinical presentation and course pose certain problems for epidemiologic investigations. Strategies for overcoming these difficulties are discussed.