It is sometimes impossible to arrest all venous oozing from raw surfaces, and the blood must be left for absorption by the peritoneum, or must be carried off by drainage with the glass tube or with gauze. Drainage enables the operator to watch the amount of hemorrhage after operations, so that if excessive he may employ measures to check it. Drainage also acts as a hemostatic. The direct pressure of the gauze upon the bleeding area checks the hemorrhage, and the continual removal of blood, the promotion of dryness, and the contact of air through the glass tube have a decided hemostatic effect. Drainage, therefore, is sometimes used not only to remove blood, but to aid in arresting hemorrhage. As the operator becomes more experienced he practises more perfect hemostasis, and learns to obliterate by buried suture, to fold in, or to cover with peritoneum raw bleeding surfaces, so that drainage as a means of hemostasis is less often required. If the operator fears that the peritoneum has become infected from imperfect asepsis at the operation, or from the escape into it of some septic material like pus, he should employ drainage, especially if he expects much subsequent serous or bloody discharge to take place. If the intestinal wall has been extensively injured, as we sometimes find after an adherent intestine has been liberated, drainage should be employed; for septic organisms most readily pass through such an injured wall, and the damage may be so great that necrosis may take place, with the escape of intestinal contents. It must be remembered that all purulent accumulations in the abdomen and pelvis are not septic. Such accumulations were septic in the beginning, but in the majority of chronic cases the septic organisms have died and disappeared, and the pus is perfectly sterile and harmless to the peritoneum. Consequently, if an ovarian or a tubal abscess ruptures during removal, and the contents escape into the peritoneum, drainage is not necessarily required. For a period of three years the writer had in such cases immediate bacteriological examination of the pus made, and determined drainage from the result of such examination. In the majority of cases the pus was sterile and drainage was not employed. It has been found, as would be expected, that the pus is most often septic in the cases of recent suppuration and in the chronic cases during an acute attack. Experience also teaches that suppurating dermoids are very likely to be septic. It will be seen from these considerations that in determining the question of drainage much must be left to the judgment and the experience of the operator.