The operation most in favour consists in ligation of the artery above and below the varix, the vein remaining untouched. Even this operation, however, in two cases of femoral varix failed to effect more than a temporary cessation of the symptoms, although the ligatures were placed but a short distance from the communication. Failure is due to the presence of collateral branches, which are not easy of detection. Even when the vessels lie exposed, the even distribution of the thrill renders determination of the exact point of communication difficult, and the difficulty is augmented by the temporary arrest of the thrill following the application of a proximal ligature to the artery. A successful case is reported by Deputy Inspector-General H. T. Cox, R.N., in which the ligatures were placed 1/2 an inch from the point of communication.[16] Single ligation, or proximal ligature, is useless. If the vein cannot be spared, excision of a limited part of both vessels may be preferable, particularly in those of the upper extremity. Proximal ligation of the artery combined with double ligature of the vein, as adopted in case 15 by Colonel Lewtas for a varicose aneurism, might offer advantages in some situations. Given suitable surroundings and certain diagnosis, the ideal treatment of this condition, as of the next, is preventive--_i.e._ primary ligation of the wounded artery. Many difficulties, however, lie in the way of this beyond mere unsatisfactory surroundings. It suffices to mention the two chief: uncertainty as to the vessel wounded, and the necessity of always ligaturing the vein as well as the artery in a limb often more or less dissected up by extravasated blood, to show that this will never be resorted to as routine treatment. _Arterio-venous aneurism._--Many of the remarks in the last section find equal application here, but in the presence of an aneurismal sac non-intervention is rarely possible or advisable. In the early stages the proper treatment in any case consists in placing the patient in as complete a condition of rest as possible, and affording local support to the limb by a splint, preferably a removable plaster-of-Paris case. Should no further extension, or, what is more likely, should contraction and diminution occur, it will be well to continue this treatment for some weeks at least.