In searching for a solution of the problem it is essential to have the following considerations clearly in mind. Every patient with influenza must be considered a potential source of pneumococcus or hemolytic streptococcus infection for his neighbor until he is proved otherwise by bacteriologic examination. Every person engaged in the care of patients with respiratory diseases must also be regarded as a potential source of danger. Pneumonia cannot be regarded as one disease but must be looked upon as a group of different diseases, with more or less similar physical signs and symptoms, it is true, but caused by a considerable variety of bacteria, infection with any one of which not only provides no protection against infection with another, but may even render the individual more susceptible to secondary infection. Therefore, every patient with pneumonia must be regarded as an actual source of danger to his neighbor, at least until it is established that he has the same type of infection. All these considerations are especially true in the presence of influenza, for it has become evident that many organisms readily gain access to the lung and produce pneumonia in patients with influenza which under ordinary circumstances fail to cause disease of the respiratory organs. Since secondary infection in respiratory disease is undoubtedly spread in large part by droplet and contact infection, the prevention of secondary infection must depend upon the elimination of these methods of transmission. Three solutions present themselves: (1) Ward treatment with absolute elimination of overcrowding and much wider separation of patients than has hitherto been deemed necessary; (2) segregation of patients according to type of bacterial infection; (3) effective individual isolation of every patient.