"The first incision will have divided the sphincter, but the other two will only have divided tissue external to it. Should the external apertures be so placed that a straight line drawn from the one to the other would pass behind the anus (Fig. 40), the steps of the operation could be reversed, and a director be passed in at one external orifice and out at the other, and the tissues divided. Now pass the director from the wound in the middle line into the bowel, through the internal opening, and slit up the tissue with the included sphincter. In this way the incisions will be found to be more or less T-shaped, the stem corresponding to the dorsal cut." [Illustration: Fig. 42—A diagram of severe horseshoe fistula, with five external openings.] [Illustration: Fig. 43—Diagram showing incisions necessary for the cure of foregoing with one division of sphincter.] TREATMENT OF HEMORRHAGE.—There is seldom much hemorrhage after an operation for fistula, but in some cases it may be necessary to ligate a large vessel which has been divided. If there should be a profuse general oozing, the sinus may be packed with iodoform gauze, or, if necessary, the rectum may be plugged; for this purpose Allingham ties a double string into the center of a large bell-shaped sponge, which is passed into the bowel so as to prevent the blood from escaping upward into the colon. He then firmly packs the parts below with cotton dusted with powdered alum or persulphate of iron. In order to allow the escape of flatus, a catheter may be passed through the sponge. As a rule, all hemorrhages following rectal operations are easily controlled by mild measures, such as the local application of hot water, of ice, or of some mild astringent. THE AFTER-TREATMENT.—After the operation for fistula in ano, the wound should be packed with iodoform gauze, which is left undisturbed for twenty-four hours. This is done to prevent subsequent hemorrhage. A pad of gauze and cotton and a T-bandage are next applied. The subsequent dressing of the case should be daily attended to by the surgeon himself. The parts should be kept perfectly clean, and the wound syringed with peroxide of hydrogen, carbolic acid solution, etc., after which a single piece of iodoform gauze laid between the cut surfaces of the wound will be all the dressing required. In the after-treatment of these cases I have seen the healing process greatly retarded by excessive packing of the wound with the lint, or delayed by the undue use of the probe. Such interference is to be avoided. If the granulations are sluggish, and the discharge is thin and serous, it will be well to apply some stimulating lotion, such as peroxide of hydrogen or a weak solution of copper sulphate (two grains to the ounce).