It is easily obtained by staining a cover-glass with a blood smear in a gummy solution of iodine and potassium iodide. When the blood is normal all the cells take on a uniform, bright-yellow color, while the white cells stain more lightly than does their protoplasm. When the glycogen reaction is present, brown granules are seen in the protoplasm of the polynuclear leukocytes, which may often take on a different brown tint. Frequently brown particles are to be seen outside of the corpuscles, while occasionally the other forms of leukocytes show also the reaction. The value of a careful blood examination is well illustrated by Plate I, prepared by Dr. Irving P. Lyon, in which are displayed the alterations of greatest interest to the surgeon. HEMOGLOBIN. The principal interest of the red blood corpuscles for the surgeon, aside from their relative number and shape, inheres in their relation to hemoglobin, and hemoglobin is of particular interest here because much can be learned by estimating the proportion in which it is present. Hemoglobin has, furthermore, an interest which reaches beyond the mere blood appearance, since it is considered to be the apparent source from which both the urinary and biliary pigments are produced. That the amount contained in the blood varies within wide limits under different conditions has long been known. When notably reduced in amount the condition is referred to as _oligochromemia_. The ideal normal standard is present in but a small proportion of cases, even in strong young men in the third decade of life. The average is considerably lower and can scarcely be placed above 90 per cent. Females show a smaller amount than males--3 or 4 per cent. less. In anemia its reduction is not usually proportionate to that in the number of red cells. After hemoglobin loss, as after surgical operations, much can be gained in the matter of prognosis by estimating the speed of its regeneration. With regard to how much actual hemoglobin loss a patient can bear, it seems to be more important to determine how much still remains in the body. The minimum is apparently 20 per cent. In three cases dying of collapse after operation, Mikulicz found only 15 per cent. remaining. The rapidity of regeneration is a fairly accurate indication of improvement in every other respect. Regeneration is interfered with by constitutional syphilis, and, on the other hand, is often apparently favored in cases of tuberculosis. In malignant tumors the average of hemoglobin is reduced to about 60 per cent., and in these cases also complete regeneration is materially retarded. Incomplete removal or recurrence of cancer prevents typical regeneration or restoration, while, after successful or radical removal, complete restoration to the previous standard, often with positive gain, is obtained. Thus, a woman who had gained thirty pounds after resection of a cancerous pylorus, showed, after three months, hemoglobin repair to the amount of 65 per cent.