=Direct Transfusion.=--The obvious method of performing a direct transfusion is by making an end-to-end anastomosis between an artery of the donor and a vein of the recipient. The most readily accessible artery is the radial at the wrist, and this is indeed almost the only artery that is available. The most accessible vein is the median basilic or the median cephalic at the elbow. The operation of end-to-end anastomosis, using an artery of so small a calibre as the radial artery at the wrist is usually found to be, is one of great technical difficulty; this effectually prevented transfusion from being used at all frequently. A modification has been used by Sauerbruch and others, in which the end of the radial artery is drawn into the lumen of the vein through a slit in its wall. A suture is passed through the radial artery close to its cut end, and the needle is then passed through the slit in the vein and out again through the wall of the vein an inch or so higher up. Traction on the suture then pulls the artery into the vein. The artery has meanwhile been temporarily occluded by a clip, which is removed when the artery is inside the vein, so that the blood can then flow from one to the other. This is easier to do than the anastomosis, but, in addition to the other objections to direct transfusion to be mentioned presently, the difficulty occurs of occlusion of the artery by the physiological process of inversion of its coats at the cut end. This is likely to happen before much blood has passed, so that apparent success at first is often not maintained. Sauerbruch claimed that the amount of blood that had passed could be estimated by measuring the time taken for 1 cc. of blood to flow from the artery before it was introduced into the vein; but there is no proof that the rate of flow remains constant. If direct transfusion be desired, there can be no doubt that Crile’s method, introduced some fifteen years ago, is the best to employ. After much patient work Crile perfected a method of anastomosis which ensures that no occlusion of the vessels can take place at the site of junction. This depends on the use of a short silver tube, through which the end of the artery is threaded. The artery is then pulled back again outside the tube in the form of a cuff and fixed in position. The end of the artery has thus been made rigid, and over this the vein is pulled in its turn and fixed by a ligature. A watertight junction is thus made, and blood can flow through it without interruption--unless clotting takes place in the vessels as the result of handling and injury to their walls. This method has been extensively used in America, and it was the first to render the operation of transfusion a comparatively popular one.