Abstract
1. Type III antipneumococcus serum, after absorption with the specific carbohydrate, no longer forms a precipitate with the carbohydrate, but still has a definite, though diminished bactericidal action on virulent pneumococci in a bactericidal test. 2. Such an absorbed antiserum still retains some of its power to neutralize the antibactericidal effect of the specific carbohydrate in a bactericidal test, showing that absorption with the carbohydrate does not remove all the anticarbohydrate antibody from an antiserum. 3. This carbohydrate neutralization test is a very much more delicate method for detecting the anticarbohydrate antibody (precipitin) than the precipitin test. 4. There is therefore no necessity to predicate another antibody to explain the bactericidal action of a carbohydrate-absorbed antiserum, or a similar result in a mouse protection test. 5. The specific carbohydrate has a definite antibactericidal action, but it is demonstrated that, were it present in this form in the body during pneumonia, it could not conceivably be produced in sufficient quantity to influence the disease.