1. There is in the filtrate of a 5 day broth culture of Type III pneumococcus a type-specific substance which has a very powerful antibactericidal action. If the precipitinogen content of the broth filtrate and the specific carbohydrate is taken as the basis of comparison, it requires approximately one thousand times as much antiserum to neutralize the broth filtrate as is necessary to neutralize the specific carbohydrate. The active substance in the broth filtrate appears to be related to the specific carbohydrate. Its possible nature is discussed.
2. A similar substance, but in stronger concentration, was found in the filtrate of a lung from a Type III pneumonia autopsy. The influence of this substance on the disease is discussed.
3. One specimen of Type III convalescent blood, though comparatively weak in anticarbohydrate antibody (precipitin) was better able to neutralize the broth filtrate and the lung filtrate than a corresponding mixture of normal blood and antiserum. Two other specimens of Type II convalescent blood neutralized the Type III broth filtrate efficiently.