A few years ago a transfusion service was organized at the Milwaukee Children's Hospital by one of us (S. J. S.) of the surgical staff. Before that time transfusions were done by the resident physician or the interns, and often the only record of the operation was a terse statement on the hospital chart that the transfusion had been made. Since we took over the service, one man has been responsible for all of the transfusions performed on patients in the hospital, and accurate records have been kept. Necessarily a great many blood groupings and compatibility tests have been made preliminary to the transfusions. It occurred to us that a study of this material might be of value, not only in giving us a better understanding of the fundamental principles involved, but also as a contribution to the studies of hereditary and forensic problems that are being discussed in the