This cooperative study was performed in the provinces of Bologna and Ferrara to estimate the prevalence rate of Myasthenia Gravis (MG). Two independent groups used the same case collection method and diagnostic criteria. The study was performed from January 1, 1987 to March 1, 1988. Cases were collected through: 1) review of clinical files from neurology clinics in Bologna and Ferrara, and from the main North-Italian centers for MG (neurology clinics of Milan and Turin; thoracic surgery unit, Turin); 2) thoracic surgery and intensive care units in Bologna and Ferrara; 3) physicians prescribing anticholinesterase drugs. Prevalence rates of generalized MG per million population were 60.9 in the city and 41.0 in the remaining province for Bologna and 118.9 in the city and 87.9 in the remaining province for Ferrara. The difference in prevalence rates between the two areas was statistically significant. Our survey shows that prevalence studies on MG may be biased by factors that cannot be controlled and that most previous studies have probably underestimated the prevalence of MG.