Trousseau,1in 1856, noted that thrombosis of peripheral veins may be the first clinical indication of the presence of a malignant tumor in some part of the body. This observation has had such recent confirmation as the study of Woolling and Shick.2They reported 15 cases of carcinoma that had been unrecognized until after the patients sought medical aid because of symptoms referable to thrombophlebitis. The association of venous thrombosis and carcinoma of the pancreas was carefully investigated in autopsy material by Sproul,3who found thrombosis of one or more veins in 14 of 47 patients with carcinoma of the pancreas and in 9 of the 16 patients with carcinoma restricted to the body or tail—a very high incidence. Generations of medical students have been taught, on the basis of these and similar studies, that peripheral thrombophlebitis as a presenting complaint is suggestive of malignant disease in