91 patients with chronic symptoms of polyneuropathy without any previously known underlying disorder underwent extensive clinical and laboratory investigations in an attempt to reveal the cause of the polyneuropathy. A definite or probable origin of polyneuropathy was detected in only 8 patients (9%) and a possible but questionable cause was found in 16 (17%), while the cause in 74% remained undetermined. In only 1 patient could effective treatment be instituted. No evidence was found for an association between polyneuropathy and reduced glucose tolerance without manifest diabetes mellitus. The findings suggest that only a limited investigation is justified in most cases of chronic polyneuropathy without obvious cause.