Various other devices for achieving the same result have been elaborated by other workers, and attention may be drawn to those of Elsberg and Bernheim, both of which are described in the book by the latter on “Blood Transfusion.” During the war a simpler method was introduced by Colonel Andrew Fullerton, who, working at a Base Hospital in France, found that he could get good results by employing a thin rubber tube with a small silver cannula at either end. The apparatus was first coated on the inside with a thin layer of paraffin wax, in order to discourage clotting within the tube, and the cannulæ were introduced into the donor’s artery and the recipient’s vein respectively. The blood could then flow freely from one to the other. The fact that blood was being transmitted was taken to be proved by the visible pulsation of the thin rubber connecting-tube synchronously with the arterial pulsations. The disappearance of this was assumed to be evidence that clotting had occurred. This method was described by Colonel Fullerton to the surgeons working at the Casualty Clearing Stations, where blood transfusion was likely to be of most service, but it was never used extensively. The coating of the inside of the tube with paraffin is in itself an operation of some difficulty. Under conditions in which any loss of time could not be permitted, success by this method was not attained with sufficient certainty, and it was shortly afterwards replaced by the more satisfactory methods described below. The most recent work on direct transfusion has been done by J. M. Graham at Edinburgh, who has however reached the conclusion that the technique is always more difficult than that of indirect transfusion. It can easily be seen, therefore, that all the known methods of direct blood transfusion present great technical difficulty, which renders the method unsuitable for general use. There are, in addition, certain other objections to it of an obvious nature. It is, in the first place, impossible to measure the amount of blood which has passed from the donor to the recipient. Sometimes an indication may be obtained from the evident improvement in the condition of the patient, accompanied by the signs of loss of blood in the donor. More often clotting takes place, unknown to the operator, at some point, with the result that blood ceases to pass a considerable time before the end of the operation, and the patient has consequently received very much less blood than is supposed. It has been claimed by Libman and Ottenberg that the amount of blood transferred may be estimated by weighing the donor before and after the operation. This presupposes that a very accurate weighing machine is easily available, which usually is not the case.