It is improbable that notching of the two vessels can primarily produce a pure varix, although it may result in the formation of an arterio-venous aneurism, especially if the bullet should have passed between the two vessels in such a way as to notch the contiguous sides. It is impossible to say, in any given case, what the result of secondary contraction of a sac produced in this manner may be in the determination of the ultimate relation of the vessels. In many of the cases clinically designated pure varix, the remains of such a sac may still actually persist. In the case also of pure perforation of the vessels, it is difficult to believe that a localised blood cavity has not originally existed. Given complete division of the vessels, as far as my experience went, arterial haematoma was the uniform result. Under these circumstances I am inclined to believe that a symmetrical perforation of both vessels is the most common precursor of either condition; that the pure varix is the rarer and less likely result, and that its formation is dependent mainly on certain anatomical conditions. The most important of these conditions are the proximity and degree of cohesion of the two vessels, the comparative spaciousness or the opposite of the vascular cleft, and the degree of support afforded by surrounding structures. Thus, the close proximity of the popliteal artery and vein, together with the particularly firm adhesion which exists between the vessels, probably favours the formation of a varix; again, a varix more readily forms if the femoral artery and vein are wounded in Hunter's canal than if the injury is situated high in Scarpa's triangle, where the vessels lie in a large areolar space. The passage of a bullet between an artery and vein may perhaps produce either condition, but wide separation of the two vessels, as for instance of the subclavian artery and vein, renders an aneurismal sac almost a certainty. These suggestions seem borne out by the cases recounted below, since the pure varices are one femoral, one popliteal, and one axillary. I cannot include the calf and forearm cases, as the existence of a small sac could not be disproved. To these anatomical factors certain others must be added. In most cases a false sac exists at first, which tends to undergo contraction and spontaneous cure, as is observed in some of the ordinary traumatic sacs. This history of development is moreover supported by the observation that proximal ligature of the artery usually converts an arterio-venous aneurism into an aneurismal varix. The process is no doubt favoured by cleanness and small size of the perforation, moderation in the amount of primary haemorrhage, the tone and resistance of the surrounding tissues, special points in the circulatory force and condition of the blood, and the possibility of maintaining the part at rest after the injury.