appear as bottles, with narrow necks and wide bottoms, filled with cells staining more intensely than normal gland-cells with carmin.

In certain parts of the tumor mass in the mucosa there occur carcinomatous bodies so arranged as to look like hypertrophied gastric glands, giving the tumor the structure of an adenoma. In other places no distinct architecture is followed, but irregular figures, now round, and then irregular, with spikes or branches on the sides, occur, thus deviating from the structure of adenoma.

Cornil thought the dilatation of the lower ends of the gastric glands was due to the following changes: the surface villi or papillæ were lengthened, and it was impossible to find the point of exit of the dilated glands at their bases. From this he concluded that, as the papillæ grew in length, they also grew together at their bases, thus closing the mouths. of the glands, and this led to a dilatation of the fundi below. This dilatation he believes to be the first step toward malignancy, and Waldeyer agrees with him here.

Granulation tissue (“periacinöse Wucherung") surrounds the cancerous bodies. This is young connective tissue, and, according to Waldeyer, the amount and consistence of this stroma decides whether a tumor is to be of the scirrhous or medullary type. Waldeyer's descriptions of the development of carcinoma from the gastric glands are plain, indeed, but his illustrations fail to show the connection between the glands and the tumor in the submucosa.

As an illustration of the development of carcinoma from the glands I offer the following case:

CASE VII.-Mixed tumor. “Cylinder-cancroid” and carcinoma. Ring-shaped, large, semisolid, ulcerated tumor in pyloric end and the neighboring part of the duodenum. Some contraction of pylorus without dilatation of the stomach. The tumor started in the glands of the mucosa, which sends extensions down into the muscularis mucosa. Similar tumors of the lymph-glands of the lesser curvature of the stomach; cylinder-cancroid of the liver. Miliary carcinoma of the peritoneum. Croupous pneumonia of the middle lobe of the left lung. Bronchiectatic cavities of the basal lobe of the left lung.

Henriette Jörgensen, forty-six years old, married. Autopsy, March 5, 1872, seventeen hours after death. The abdominal cavity contains 2.5 liters of a clear, pale, serous fluid. The pylorus is pulled down and lies below the right lobe of the liver, surrounded by nodular tumors. The stomach is not dilated; it holds 520 c.c. The form is normal as far as to the pyloric end, which is contracted and does not admit the tip of a finger. The constriction is due to an ulcerated tumor which is 10 cm. wide and reaches completely around the pyloric end. The tumor ends toward the fundus as a pronounced infiltrated, wave-shaped ridge. It extends over the pyloric sphincter for 2 cm. into the duodenum as a flat, grayishwhite, soft infiltration of the mucosa.

Cross-section at the pylorus shows the serosa on the outside, and is unchanged except at the lesser curvature, where there is found a tumor mass, 1 to 2 cm. in thickness; the muscularis is hypertrophied, being 5 to 10 mm. thick, and penetrated by large septum extending into it from the mucosa; in the submucosa inside of the muscularis is a homogeneous, soft, white tumor which in cross-section simulates the brain substance. From the cut surface a purulent fluid, consisting of cylindric cells, can be expressed. The mucosa of the fundus has undergone postmortem softening. In the gastrohepatic ligament occur small, firm, white nodules. Everywhere around the pylorus are tumors varying in size

from that of a hempseed to a walnut. One of the retroperitoneal glands has been changed into a soft, walnut-sized tumor.

At the upper right edge of the right lobe of the liver is found a round, nodular tumor as large as a walnut. Besides this tumor there occur a nodule as large as a hempseed and another as large as a pea in the left and right lobes of the liver respectively.

Tumors varying in size from a millet-seed to that of a pea are found in the uterovesicular fossa and in the pouch of Douglas. These rather firm tumors are fastened to the other tissues by a broad base. They are often pigmented, and have a papillomatous surface. The other organs are unchanged.

Microscopically, the mucosa is seen to extend 1 cm. up over the edge of the ulcer. The edge is 1 to 2 cm. thick, and consists of a 1⁄2 to 1 cm. thick layer of carcinomatous cells, and below this the thickened muscularis, which is penetrated by white septa. Underneath the muscularis, in the neighborhood of the lesser omentum, is found a tumor mass 2 cm. thick, which incloses a number of enlarged lymph-glands. The mucosa at the edge of the ulcer is not thickened; on its surface occur naked villi which are not enlarged; no exits for the gastric glands are visible, and the mucosa consists of a uniform mass of connective tissue containing scattered indifferent cells. In the deeper layers of the mucosa are glands occurring as distinct groups of acini. A few of these are rather winding in their course, and have slightly irregular sides. In the muscularis mucosæ appear alveoli, some of which are regular and lined with cylindric cells; others are irregular, often with long, narrow cavities, which are filled with round polymorphous or cubic cells, 12 micra large, and having a nucleus 6 micra in size. These cavities-Waldeyer's carcinomatous bodies— are clearly seen to be in direct connection with, and to originate in, the glands of the mucosa (See Plate I, Fig. 2, 5). The tumor in the submucosa consists of a connective-tissue stroma containing alveoli lined with cylindric cells. These alveoli are often arranged in groups resembling the grapelike glands of the mucosa. In the white septa between the muscular bundles alveoli similar to those of the tumor occur. In the lymph-glands is found a delicate, often pronouncedly degenerated, stroma, surrounding large groups of alveoli, most of which are lined with cylindric cells, but some are filled with irregular cell forms. Large parts of the lymph-glands consist of cheesy masses in which the outlines of regularly arranged cylindric cells can be distinguished in a fatty, granular detritus.

The tumors in the liver are cylindric carcinomata having a scant stroma and large alveoli.

As is seen in Fig. 2, Plate I, the connection between the gastric glands and the carcinoma is definite. The papillæ at the surface of the mucosa are seen to be of normal height; their epithelial covering is absent. The gland openings are not visible, and the upper layer of the mucosa appears as a uniform layer of connective tissue containing fibers and nuclei partly arranged along the long axes of the villi.

It was pictures of this type which brought Cornil to the conclusion that the mouths of the gastric glands were closed by the growing together of the villi at the bases. I consider this conclusion rather hasty, since the identical picture can be seen in the upper layer of the mucosa in cases in which the epithelium is lacking from the villi and the openings of the glands; the openings become unrecognizable in such cases, but may be wide open for all that. That such is the case I have often observed while examining the mucosa of parts of the stomach distant from the malignant growth, as, for instance, in the fundus, when the cancer was located at the pylorus. Whether the epithelium in such cases was lost prior to death through catarrh of the stomach or was lost by softening after death, one gets the impression of the upper part of the mucosa

being a homogeneous mass with closed gland mouths. But this does not mean that they are actually grown together, especially since the fundi below are of normal appearance, neither filled with cells and secretion nor dilated, hence without any signs of anything preventing their contents from being emptied. Deeper in the mucosa (See Plate I, Fig. 2) the glands (3) appear as groups of acini cut across. These are not dilated, but here and there they are slightly irregular in form. This is especially the case in one place (at 5, a), where they break through the muscularis to the mucosa; here occurs a group of irregular acini, from which a string of cancer-cells penetrate into the muscularis mucosæ (5, a). This string is at first narrow, being only two cells wide, but a little deeper where it enters the submucosa; branches are given off at the sides, forming elongated groups of cancer bodies (5, b). The entire group of cancerous bodies and glands correspond nicely to Waldeyer's description; the group has the shape of an hour-glass, the upper, wider portion being formed by the irregular acini, the middle part by the narrow stalk in the muscularis mucosa, and the lower, wider portion (5, b) by a larger group of cancerous tissue in the submucosa. A little to the left of the figure (at 5, c) is seen another penetration; in the narrow septum in the muscularis occurs a canal cut across, lined with regularly placed cells; below this is seen both a large cancerous body, consisting of roundcells, and a group of alveoli similar to the irregular acini of a gland, having regular secreting cells in one of the acini. The connection between these cancerous masses and glands above is not shown in the figure, but such a connection was without doubt present. It is, as a rule, very seldom, that the origin from the glands is shown by a single section as plainly as at 5, a. In such a case the cut must pass straight in the direction of the penetrating cells, for if it falls in any angle of this, isolated groups of alveoli appear; when, however, one has once discovered the direct connection between glands and the carcinomatous tissue underneath, it is easy to follow, by successive sections, the relationship of the malignant cells, although the sections may run diagonally to the stalk.

RETROGRESSIVE PROCESSES AND ULCERATION IN CANCER OF THE

STOMACH

On the whole, the retrogressive processes of carcinoma of the stomach are similar to those of carcinomata anywhere in the body. When cancerous tissue has reached a standstill, there occurs, after a certain period, different for the various types, and, as a rule, shorter for the softer than the more solid forms, fatty degeneration, whereby both the stroma and the alveoli are changed into a granular detritus. In carcinoma of the stomach, and especially in the softer forms, as medullary carcinomata, large portions of tissue may die at once, thus producing greater or smaller areas of gangrene in the tumor.

(A) Gangrene.-Gangrene of larger portions of malignant tumors is not a rare occurrence. Prus noted it, and believed it might lead to a

*Loc. cit., p. 282.

cure. Köhler states that the entire tumor might become gangrenous, and the process may not only extend to the serous coat of the stomach, but might affect the adherent neighboring organs as well. Lebert* once observed a gangrenous slough adherent to the liver. Later observers, though reporting large series of cases, do not put any special emphasis on gangrenous destruction. For this reason I shall describe a case of this type. In this case carcinoma was suspected during life, since the vomited material had a pronounced fetid odor.

CASE VIII.-Medullary carcinoma (medium-sized cells). A soft ulcerated tumor in the pyloric end. This is adherent to a tumor as large as a child's head between the pylorus and the transverse colon. The center of the large tumor is gangrenous, and a large carity has been formed which communicates with the stomach through the bottom of an ulcer. Carcinomatous lymph-glands along the lesser curvature. Marantic thrombi of both crural veins.

Marie Richter, fifty-nine, married. Autopsy July 9, 1873, thirty-six hours after death. There is no fluid in the peritoneal cavity. Between the transverse colon and the stomach is a tumor as large as a child's head. It is found in the left hypochondrium and the iliac fossa, and it is movable against the posterior abdominal wall, but adherent to the stomach and the colon. The stomach contains a thin, grayish fluid in which are whitishgray shreds and clumps of degenerated tissue. In the posterior lower wall of the pylorus is a round ulcer, 10 cm. in diameter. This has a gray spotted, soft, uneven raised edge. The floor forms the entrance to a cavity as large as a fist. This has been produced by gangrene in the large tumor, and is filled with a gray, stinking fluid containing large necrotic masses of the malignant tissue. The non-gangrenous part of the tumor has a grayishwhite cut surface, from which great quantities of purulent fluid can be expressed. This fluid contains polymorphous cells, part of which are cylindric and part round or angular, with round or oval nuclei. The surface of the tumor is covered by the omentum, and here are found numerous soft white nodules varying in size from that of a pea to a nut. There are metastatic tumors along the lesser curvature. No other metastatic growths are present.

(B) Ulceration.-Gangrene of smaller portions of the tumors is often associated with ulceration, so that there is found in the rapidly growing tumors a combination of gangrene and molecular necrosis. (Rokitansky and Dittrich were the first to write detailed descriptions of ulceration in carcinoma of the stomach.) In the firmer types, as the diffuse scirrhous and alveolar carcinomata, retrogressive processes occur from the surface inward and take place so evenly that the stomach-wall appears ground away. Dittrich believes this might be due to the peristaltic movements of the organ, whereby the tumor is rubbed against the food or the opposite wall.

There are no statistics giving the frequency of ulceration for the various types of carcinomata. The statistics available give only the relative proportions of ulceration in carcinoma as a whole. All reports are vague regarding the frequency of ulceration in any individual class of cancers. Köhler, who grouped together the figures given by Lebert, Valleix, and Dittrich, found ulceration present in 69 of 115 cases, or 60 per cent., of carcinomata of the stomach. Waldeyer found 32 cases, which he divided as follows: 11 scirrhous, 5 medullary, 3 colloid, and 13 *Loc. cit., p. 483.

ulcerated. It cannot be possible that more than half of Waldeyer's cases were non-ulcerative, and just what he means by 13 ulcerative cases is hard to say. Twenty-eight of my 32 cases were complicated by ulcers, and only 4 tumors-2 of diffuse scirrhous, 1 of cylindric celled, and 1 of diffuse alveolar carcinoma-were covered with intact mucous membrane, which never appeared quite normal, and was at times more or less infiltrated with carcinomatous cells. There are, of course, all possible varieties of malignant ulcers, from the most superficial ones to a very wide and deep loss of substance, and to place an arbitrary limit on it, as has been attempted by Köhler and Waldeyer, is impossible.

The ulceration begins in the mucosa, which either sloughs off in large flakes or disappears in very small particles. As far as I can make out, from a study of younger tumors, the degeneration of the mucosa is always due to a preceding cancerous infiltration, hence it is the carcinomatous tissue which becomes necrotic. I have never observed that any large part of the non-cancerous mucosa has sloughed. Destruction of the malignant tissue in the submucosa and the muscular layer is often accompanied with small hemorrhages. The blood when changed by the gastric juice appears like coffee-grounds in the vomitus. The degenerated tissue is also infiltrated by this blood, so that the floor of the ulcer often has a grayish-black appearance. The destruction of tissue does not, as a rule, take place evenly over the entire surface, but often penetrates into the depths of the tissues, forming irregular cavities.

While destruction of tissue takes place in some places, new cancerous projections grow out in others, thus adding still more to the irregularity of the ulcer. One remarkable feature of these carcinomatous ulcers is that there is rarely any disagreeable odor from them. Neither the vomited masses nor the gas forced up from the stomach is especially odorous. This is apparently due to the action of the gastric juice, since it has lately been shown to have a deodorizing effect on cancers in other parts of the body. As far as the gastric juice goes, it is assumed to have some effect on the ulcerating surface which is not protected by the mucosa. It seems to dissolve the superficial layers of the surface it touches, and thus facilitates the spreading of the ulcer. In a few instances I have observed what appeared to be a dissolving action of the gastric juice on the edge of the penetrated muscularis. I found, for instance (Case XXVIII), that the infiltrated muscular edge of the ulcer was smooth. Microscopically, the carcinomatous tissue and the musclefibers did not show fatty degeneration, as is usually the case in carcinomatous ulcers, but they appeared unchanged almost to the very edge of the ulcer. There was only a thin layer (0.5 mm. at the very edge) where the malignant tissue, as well as the muscle-fibers, was changed to a finely granular mass without fat-globules. I believe it possible for this finely granular degeneration to be due to the corroding influence of the gastric juice. This is made more plausible by the fact that such a finely granular edge has been observed* in the peptic ulcers, and here it is recognized as due to the action of gastric juice.

* Klebs: Loc. cit., 1. Lieferung, p. 182.