Many authorities refuse to regard purpura as a disease entity and consider it merely as a manifestation of some other disease. Its extreme rarity is evidenced by the fact that at the Massachusetts General Hospital there were only sixty-four cases in 155,884 medical and surgical admissions; at Johns Hopkins Hospital there were forty-one cases in 18,594 medical cases, and at the Hamburg General Hospital there were seventy-three cases in 100,000 admissions; B. Ramwell1 saw sixteen cases among 5,256 ward patients. In his table of 258 cases of all kinds of purpura fifty-four were primary and showed arthritis. In looking over the literature of purpura, it is difficult to find any mention of familial tendencies in this disease. In fact, familial tendencies have been used as a point against purpura in a differential diagnosis from hemophilia, so rarely are they found. There are, however, a few instances reported to which we