Chart 1—Showing the relation of (_a_) onset of cases of pneumonia shown by autopsy to be uncomplicated by secondary infection with hemolytic streptococcus, indicated by upper continuous line with single hatch, and of (_b_) onset of fatal cases of streptococcus pneumonia, indicated by the lower continuous line with double hatch, to (_c_) the occurrence of influenza, indicated by the broken line. The onset of each case of fatal pneumonia is represented by a single square. ] Our study of ward infection in pneumonia furnishes an explanation of the outbreak of fatal streptococcus pneumonia coincident with the initial stage of the influenza epidemic. This outbreak is a true epidemic of streptococcus infection superimposed, in many instances at least, upon preexisting pneumococcus pneumonia, but in some instances, doubtless, a primary streptococcus pneumonia, following the bronchitis of influenza. In the absence of secondary streptococcus infection a very large proportion of these individuals would have recovered. This epidemic of streptococcus pneumonia, it has been shown, was the result of unfavorable conditions produced by great overcrowding of the hospital and in the early part of the epidemic by inadequate separation of those with streptococcus infection from those with none. With control of these conditions, streptococcus pneumonia rapidly diminished. Greater susceptibility to pneumococcus pneumonia in the early than in the late period of the epidemic is perhaps explained by differences in the severity of influenza; the more susceptible individuals were attacked by influenza in the early period, whereas the less susceptible did not acquire the disease until they had been exposed to an immensely increased number of infected individuals. A better explanation is furnished by the greater opportunity at the beginning of the epidemic for the transmission of microorganisms causing pneumonia, for at this time patients with influenza were crowded together and methods to prevent the transmission of infection were little used. Bronchitis Clinical study has shown that purulent bronchitis (see Fig. 2) occurs in about one-third of the cases of influenza. In a large proportion of cases of bronchitis there is no clinical evidence of pneumonia. The bronchial lesions found in association with the pneumonia of influenza are an index of the ability of the agent, which causes influenza, to injure the bronchi.