extension to or metastatic involvement of other organs. The finding of metastases, especially those occurring in the pouch of Douglas, may permit a differentiation between a filling defect due to ulcer and one due to carcinoma. It should be more generally recognized by röntgenologists that examination of the lower bowel may reveal evidences of the so-called " Douglas metastases," though, on account of the easy accessibility of the rectum for digital exploration, physical examination is also very useful.

In differentiating between benign cicatricial stenosis of the pylorus and stenosis due to malignancy the writer has found it of especial value to make the screen and plate examination with the patient lying on the right side, the tube behind the patient, and the screen or plate held vertically against the abdomen. In this manner it is possible to bring out the finest detail of the pyloroduodenal region, often to better advantage than with the patient in the prone position, plate anterior. Unless the pyloric carcinoma has supervened upon an old stenosing ulcer, it is likely that the stomach will not be greatly dilated in pyloric cancer, for the reason that the malignant process has advanced too rapidly to permit extensive dilatation. In benign ulcerous stenosis, on the other hand, including those cases where the ulcer has later degenerated into malignancy, the long duration of the process permits enormous increase in the size of the stomach.

Extraventricular tumors, unless intimately adherent to the stomach, may be differentiated through the fact that the normal gastric peristalsis is not interfered with. This is best determined by fluoroscopy, although the expenditure of a number of plates may afford the same information.

In differentiating between hour-glass due to ulcer and that due to carcinoma there are several important points to be observed. Ulcer and carcinoma show differences in position, length and outline of the connecting canal between the upper and lower sac, as well as in the relative size of the two sacs. The ulcer or its scar is almost always located on the lesser curvature. The writer has seen but one case of ulcer high up on the greater curvature, in that instance penetrating into the spleen. The contraction associated with lesser curvature ulcer always occurs toward the lesser curvature, the seat of the shrivelling agent. The connecting canal between the upper and the lower sac is located near the lesser curvature, and its outlines are

usually nearly smooth and regular. There is often a slight projection of the stomach shadow at the site of the ulcer, owing to excavation attending the ulcer process. In ulcer only a limited portion of the greater curvature is pulled over toward the ulcer; the narrowness seldom affects the greater curvature for more than half an inch. In carcinoma, on the other hand, although the tumor usually starts on the lesser curvature, it produces there a light space (filling defect), and the connecting channel between the upper and the lower sacs is located near the greater curvature. When the malignant hour-glass is produced by an annular carcinoma, the lumen between the two sacs occupies a median position, lying in the axis of the stomach. The length of the carcinomatous narrowing is greater than with ulcer, unless the ulcer has been attended by perigastric adhesions of considerable extent, or unless the ulceration has been multiple. The contour of the filling defect is irregular and often indistinct, because the wall of the stomach is infiltrated. In ulcer the pathological findings occur opposite the site of the filling defect, while in carcinoma there are resistance, often pain, and sometimes a palpable tumor corresponding with the filling defect.

Differentiation between the hour-glass of ulcer and of carcinoma is helped by a consideration of the relative size of the two sacs. This relative size depends upon the relative position of the hour-glass constriction to the pylorus, upon the degree of constriction, and upon the condition of the pylorus. The greater the constriction and the nearer the constriction to the pylorus, the greater will be the size of the upper sac. The size of the lower sac, which is of greatest differential diagnostic importance, depends upon the permeability of the pylorus. A large lower sac exhibiting vigorous peristaltic waves speaks for ulcer, owing to the tendency of the pylorus to abnormal spastic contraction in the presence of ulcer; whereas in carcinoma the absence of free hydrochloric acid with resulting relaxation of the pylorus does not favor development of a large lower sac. Hence we may conclude that in hour-glass stomach, when the two sacs are markedly different in size, and especially when the lower sac is small, this fact speaks for carcinoma.

Holzknecht was one of the first to call attention to the possibility of utilizing the Röntgen examination to draw conclusions as to the resectability of carcinoma of the stomach. Holzknecht is careful to

use the word resectability rather than operability, because metastases to glands can hardly ever be suspected, and adhesions are recognized only with a relative degree of certainty. According to Holzknecht and Haudek, in resectable cases the normal hook form of the stomach is usually preserved, while the types of tumor cases which are not resectable show the diagonal, short, small form of the stomach. Even very large tumors requiring a subtotal resection are usually still resectable if they preserve the hook form. Another symptom of resectability is the preservation of the distensibility of a considerable portion of the stomach. One should not neglect study of the phrenohepatic shadow and of the remainder of the gastro-intestinal tract, especially the pouch of Douglas, before reaching conclusions as to operability.

In a certain small class of cases where the clinical examination warrants a reasonable suspicion of malignancy, and where the Röntgen findings are negative, it is wise to repeat the Röntgen examination after four or five weeks. In a few cases the second examination has revealed evidences of malignancy not made out earlier; in the majority of cases the negative diagnosis will be strengthened. One of the most useful purposes of the Röntgen examination in this class of cases, as well as in cases of inoperable malignancy, is to save the patient from an unnecessary exploratory operation.

A paper on this subject would be incomplete without reference to a paper by White and Leonard, entitled "X-ray Evidence in Early and Latent Cancer of the Stomach" (Boston Medical and Surgical Journal, October 1, 1914). Their study includes 114 hospital and private cases examined within a year and a half by clinical and X-ray methods in which cancer of the stomach was suspected and found. In forty cases cancer was ultimately proved and in sixty cases ultimately ruled out. Eight cases remained doubtful. Thirtythree out of the forty cases of cancer were excluded from the paper as being well developed, leaving eighty-one cases of early, latent, or suspected cancer. These eighty-one cases finally were diagnosed as one early, six latent, sixty-two carcinoma ruled out, and twelve doubtful. The one early and six latent cases were confirmed by operation. In not a single case of the sixty-two in which carcinoma was ruled out has carcinoma been found. The twelve doubtful cases