Arteriovenous fistula was first mentioned in the literature in 1757, when William Hunter1 described the clinical features of the condition and the disturbed vascular physiology associated with it. In 1762 he reported the two cases, in both of which the condition was discovered by chance during phlebotomy, on which his description was based. A few years later Delacombe confirmed his observations at autopsy, and since that time a voluminous literature has grown up about the subject, most of which, however, concerns the traumatic or acquired variety.
The congenital variety has always been regarded as decidedly infrequent. Of four hundred and forty-seven cases reports of which were collected by Callander2 in 1920 at the instigation of Halsted, only three fell into this category, and as late as 1930 Lewis3 was able to find in the literature reports of only twenty-four cases, to which he added six of his