1. A study has been made by means of agglutination and absorption of agglutinin tests, of the antigenic relationships of 54 strains of hemolytic streptococci associated with an epidemic of puerperal fever in a large city obstetric hospital.
2. These organisms were cultured from the blood, vaginæ, peritoneal cavities and metastatic foci of patients, from the noses and throats of patients and staff members and from the peritoneum and axillary abscess of two of the nurses.
3. In preliminary agglutination tests twenty-one out of thirty-one cultures from the patients and nine out of twenty-three cultures from the hospital personnel, fell into a single agglutinative group.
4. By reciprocal agglutination and absorption of agglutinin tests, six of these strains were demonstrated to be antigenically identical. Two of these came from patients' blood cultures, three from patients' vaginal cultures and one from the nasal culture of one of the nurses. This nurse was, however, not the original carrier for she did not join the hospital staff until after the epidemic began.
5. By agglutination and complete absorption of agglutinin, sixteen more strains showed their antigenic similarity to, if not identity with, these other six strains. Four of these sixteen came from metastatic foci and two from the vaginæ of four of the five patients which had yielded from another source, the identical strains mentioned in 4. Nine came from eight other patients and one from the axillary abscess of a nurse. No reciprocal tests were possible with these sixteen strains because sera had not been prepared against them.
6. Five other strains which completely absorbed agglutinin had evidently lost their ability to agglutinate. Two of these strains were from the blood cultures and two from the vaginal cultures of two fatal cases. One was from the peritoneal cavity of a nurse.
7. Two of these five strains when injected into rabbits in the usual way for the purpose of producing agglutinating antiserum, stimulated the production of sera which agglutinated completely the six antigenically identical strains but failed to agglutinate the homologous strains used to produce the respective sera.
8. These two strains therefore retained the antigenic properties first of stimulating the agglutinating antibody in normal animals and second of uniting with the agglutinating antibody in the serum but lost the ability to bring about the second phase of agglutination—namely flocculation.
9. We have demonstrated, therefore, a more or less permanent splitting of the phenomenon of agglutination.