There is very little evidence, as shown in the literature, that special study on the blood during past influenzal epidemics has been made. A few references to alterations in the count of cells have been reported for the last epidemic (1890), but they are, as a rule, very brief statements. Cabot notes a normal leucocyte count in two-thirds of the cases, and a moderate increase in the rest. Several observers call attention to the leucopenia during the height of the disease, with a subsequent rise after the temperature has fallen to normal. According to Rieder and Herman (American Journal of Medical Science, 1893, cv. 696), the leucocytes were not increased in simple influenza, and only very slightly in the pneumonia following this disease. Herman also noticed a decline in the leucocytes in pneumonia as a fatal ending ensued. This finding was one of the few recorded for the 1890 epidemic. Emerson (Emerson Clinic Diagnosis, 1911, 558) found in influenza almost one-half of the cases showing more than 10,000 leucocytes, some even reaching 25,000. He further notes that early in the disease the count may be low, 3,000 to 5,000, but it usually rose sharply, to fall again when the temperature comes to normal. He lays stress on obtaining a leucocyte curve for each case in order to get a true picture of what changes occur. The past epidemic has brought out many observations on this subject. They vary somewhat, as is to be expected, but a common factor seems to be more or less basic—namely, a leucopenia or a normal count is the most significant single blood picture we have of uncomplicated influenza. Further, a leucocytosis is fairly generally, and we believe correctly, interpreted as evidence of a secondary bacterial invasion in this particular epidemic, and usually of the respiratory system. The leucopenia is as much a part of the clinical picture of influenza as it is of typhoid fever. Leucocytosis always means secondary invasion by other organisms. During the recent epidemic the clinical laboratory department of the School of Medicine, University of Pittsburgh, has made 747 blood counts on influenza cases. In most of the cases blood counts were made as a routine, while repeated counts were done only on selected patients.