The factors governing the reaction between chloroform and hemoglobin by which the hemoglobin of the blood may be precipitated were studied, and a process, based upon this reaction, for the separation of a clear sterile serum from old defibrinated blood antitoxin was devised.

It is shown that the yield of clear serum separated by this process approximates 70% of the original volume and that the product of the process is free from bacterial contamination. While it is shown by analysis that the globulin content of the serum suffers a slight loss, when the serum is treated by this process, potency tests indicate that the loss in antibodies that occurs during clarification may be disregarded.