The possibility that the prevalence of diseases in which there is no specific causal organism may be affected by the prevalence of other diseasses of a contagious nature is well recognised. Even in the early days of bactriology interrelationships between infectious diseases and puerperal septicaemia were suspected on purely statistical grounds. Longstaff (1891), for example, pointed out the remarkably close correlation between the seasonal variations and the secular trends of the mortalities of erysipelas and childbed fever—a relationship so close that he “found it difficult to avoid the conclusion that they were both due to the one poison.” A somewhat similar but less striking association was shown with other inflammatory diseases, such as pyaemia, scarlet fever, “rheumatism of the heart or pericardium” and diphtheria. Even more emphatic were the views of Minor (quoted by Longstaff) who, with reference to the association between erysipelas and puerperal fever, gave reasons for the belief that there existed an intimate connection between them, and that “in any place where erysipelas is found, there will be found puerperal fever.”