The regional variation in cerebrovascular disease mortality within England and Wales has been described and analyzed using data for two time periods: 1969–1973 and 1975–1979. A stable geographical pattern emerges which is substantially similar for the two sexes. The possibility of this variation being caused by differences in certification practice is discounted. The variations show some significant correlation with indices measuring the size and quality of medical services. A significant positive correlation between the geographical patterns for cerebrovascular disease and ischaemic heart disease has been found. Possible explanations are discussed. The regional variation in the level of improvement of age/sex specific mortality rates has been considered but no correlation found between the initial rates and the size of the subsequent improvem