Very early in the epidemic we realized that the pneumonia was of unusual severity and most difficult to treat satisfactorily. We were at once impressed by our helplessness, particularly in those patients showing cyanosis. Nothing we did seemed to vary the course of the pneumonia after this sign was evident. Our work in the epidemic began about October 10 on receiving a large batch of soldiers, about 100, from the Student Army Training Corps of the University of Pittsburgh. At the end of the first week several points were impressed on our mind. Firstly, in the severe cases of pneumonia; and in the early part of the epidemic most of the pneumonia was severe, the mortality was excessive, much higher than we have been accustomed to experience in Pittsburgh, where, as a rule, our hospital ward pneumonia is a very severe infection. Secondly, the wide variation in the severity of the epidemic as presented in the student soldiers coming from identical surroundings and conditions, the mildness on the one hand and the malignant character of the influenza on the other, was a very striking feature. This led to our adopting a form of treatment which was quite successful. We worked purely on the hypothesis that those individuals recovering from a mild or moderate influenza infection developed a higher grade of immunity than those in whom the disease was more severe or fatal, and this immunity could be transferred to another. This, of course, was merely inference. If the mild cases did present a higher immunity, one would naturally think that immune bodies would be present in the blood, and that in transfusion from cases which had recovered one might have a measure of therapeutic value for this epidemic. Recently Spooner, Scott and Heath and others have demonstrated specific agglutins in the serum of patients convalescing from the epidemic. On October 17 we gave whole citrated blood from a convalescent case of uncomplicated influenza to an influenzal pneumonia patient. The result in this case was strikingly good, and for the following five or six weeks this method was frequently used. We decided to give the whole blood instead of the serum, as we were able to treat the cases more readily and rapidly in this way. Our method of transfusion was, fortunately, very simple.