Profuse hæmatemesis is more common with soft cancers than with other forms. The source of profuse hemorrhage is in some large vessel eroded by the ulcerative process. The same vessels may be the source of the bleeding as have been enumerated in connection with gastric ulcer. Cancers situated near the pylorus or on the lesser curvature are the most likely to cause severe hemorrhage.

While it is true that coffee-grounds vomiting is most common in cancer, and profuse hæmatemesis is most common in ulcer of the stomach, it is important to remember that either disease may be attended by that form of hemorrhage which is most common in the other.

Dysphagia is one of the most important symptoms of cancer of the cardia. Dysphagia is sometimes one of the first symptoms to attract attention, but it may not appear until late in the disease. It is usually accompanied with painful sensations near the xiphoid cartilage or behind the sternum, or sometimes in the pharynx. The sensation of stoppage of the food is usually felt lower down than in ordinary cases of stenosis of the oesophagus. Stenosis of the cardia can be appreciated by the passage of an oesophageal bougie, but it is important to bear in mind that dysphagia may exist in cases of cancer of the cardia in which the oesophageal bougie does not reveal evidence of stenosis. Dysphagia may be a prominent symptom in cancer occupying parts of the stomach remote from the cardia.⁴⁷ The dysphagia here considered is not likely to be confounded with the difficulty in swallowing which is due to weakness or to aphthous inflammation of the throat and gullet, which often attends the last days of gastric cancer.

⁴⁷ A case in point has been reported by J. B. S. Jackson. The cancer occupied the pyloric region (American Journ. of Med. Sci., April, 1852, p. 364).

From a diagnostic point of view the presence of a tumor is the most important symptom of gastric cancer. In the absence of tumor the diagnosis of gastric cancer can rarely be made with positiveness. A tumor of the stomach can be felt in about 80 per cent. of the cases of cancer of the stomach (Brinton, Lebert). With all of its importance, it is nevertheless possible to exaggerate the diagnostic value of this symptom. It is by no means always easy to determine whether an existing tumor belongs to the stomach or not, and even if there is proved to be a tumor of the stomach, there may be difficulty in deciding whether or not it is a cancer. Many instances might be cited in which errors in these respects have been made by experienced diagnosticians. The value of tumor as a diagnostic symptom is somewhat lessened by the fact that it often does not appear until comparatively late in the disease, so that the diagnosis remains in doubt for a long time. It is to be remembered also that tumor is absent in no less than one-fifth of the cases of gastric cancer.

In order to understand in what situations cancers of the stomach are likely to produce palpable tumors, it is necessary to have in mind certain points concerning the situation and the relations of this organ.

The stomach is placed obliquely in the left hypochondrium and the epigastric regions of the abdomen, approaching the vertical more nearly than the horizontal position. The mesial plane of the body passes through the pyloric portion of the stomach, so that, according to Luschka, five-sixths of the stomach lie to the left of this plane. The most fixed part of the stomach is the cardiac orifice, which lies behind the left seventh costal cartilage, near the sternum, and is overlapped by the left extremity of the liver. The pyloric orifice lies usually in the sagittal plane passing through the right margin of the sternum, and on a level with the inner extremity of the right eighth costal cartilage. The pylorus is less fixed than the cardia. When the stomach is empty the pylorus is to be found in the median line of the body; when the stomach is greatly distended the pylorus may be pushed two and a half to three inches to the right of the median line. The pylorus is overlapped by a part of the liver, usually the lobus quadratus or the umbilical fissure. About two-thirds of the stomach lie in the left hypochondrium covered in by the ribs, and to the left and posteriorly by the spleen. The highest point of the stomach is the top of the fundus, which usually reaches to the left fifth rib. The lowest point of the stomach is in the convexity of the greater curvature to the left of the median line. The lower border of the stomach varies in position more than any other part of the organ. In the median line this border is situated on the average about midway between the base of the xiphoid cartilage and the umbilicus, but within the limits of health it may extend nearly to the umbilicus. The lesser curvature in the greater part of its course extends from the cardia downward to the left of the vertebral column and nearly parallel with it. The lesser curvature then crosses to the right side on a level with the inner extremity of the eighth rib, and in the median line lies about two and a half fingers' breadth above the lower margin of the stomach. The lesser curvature and the adjacent part of the anterior surface of the stomach are covered by the left lobe of the liver.

It follows from this description that only the lower part of the anterior surface of the stomach is in contact with the anterior abdominal walls. This part in contact with the anterior abdominal walls corresponds to a part of the body and of the pyloric region of the stomach, and belongs to the epigastric region. The remainder of the stomach is covered either by the liver or by the ribs, so that in the normal condition it cannot be explored by palpation.

It is now evident that tumors in certain parts of the stomach can be readily detected by palpation, whereas tumors in other parts of the organ can be detected only with difficulty or not at all. Cancer of the cardia cannot be felt by palpation of the abdomen unless the tumor extends down upon the body of the stomach. Cancers of the fundus, the lesser curvature, and the posterior wall of the stomach often escape detection by palpation, but if they are of large size or if the stomach becomes displaced by their growth, they may be felt. Cancerous tumors of the anterior wall or of the greater curvature are rare, but they can be detected even when of small size, unless there are special obstacles to the physical examination of the abdomen. Cancerous tumors of the pylorus can be made out by palpation in the majority of cases notwithstanding the overlapping of this part by the liver. The pyloric tumor may be so large as to project from beneath the border of the liver, or the hand may be pressed beneath this border so that the tumor can be felt, or, what is most frequently the case, the weight of the tumor or the distension of the stomach drags the pylorus downward. The pylorus may, however, be so fixed by adhesions underneath the liver, or the liver may be so enlarged, that tumors of this part cannot be reached by palpation.

The situation in which cancerous tumors of the pylorus can be felt varies considerably. The usual situation is in the lower part of the epigastric region, a little to the right of the median line, but it is almost as common for these tumors to be felt in the umbilical region, and it is not rare for them to appear to the left of the median line.⁴⁸ Brinton states that the tumor is in the umbilical region more frequently in the female sex than in the male, in consequence of the compression exercised by corsets. Occasionally pyloric cancers produce tumors in the right hypochondrium. Exceptionally, pyloric tumors have been felt as low as the iliac crest or even in the hypogastric region.

⁴⁸ According to Jackson and Tyson, pyloric cancers are felt more frequently to the left than to the right of the median line.

Cancers of the stomach do not usually attain a very large size. Sometimes they form visible protuberances. An important criterion of cancerous tumors of the stomach is their gradual increase in size by progressive growth.

The consistence of cancerous tumors of the stomach is nearly always hard, as appreciated by palpation through the abdominal walls. The surface of the tumor is usually nodulated or irregular, but exceptionally it is smooth. The tumor may be movable or not, but in the majority of cases it is rendered immovable by adhesions. Mobility of the tumor, however, does not exclude the presence of adhesions. The tumor sometimes follows the respiratory movements of the diaphragm, especially when it is adherent to this structure or to the liver, but more frequently the tumor is not affected or but slightly affected by the movements of the diaphragm. If the tumor is not fixed by adhesions, it may change its position somewhat according to the varying degrees of distension of the stomach or in consequence of pressure of intestine distended with gas or feces. In consequence of these movements or of an overlying distended colon the tumor may even disappear temporarily. It is possible that the tumor may lessen or disappear in consequence of sloughing of the growth.⁴⁹ It is not rare for a certain amount of pulsation to be communicated to the growth by the subjacent aorta. This pulsation is most common with pyloric tumors.

⁴⁹ Symptoms which have been considered as diagnostic of sloughing of stenosing cancers of the pylorus are diminution in the size of the tumor, alleviation of the vomiting, hemorrhage, replacement of obstinate constipation by diarrhoeal stools which often contain blood, increased pain after eating, and rapid progress of cachexia.

The percussion note over the tumor is usually tympanitic dulness. Sometimes there is very little alteration over the tumor of the normal tympanitic note belonging to the stomach; on the other hand, exceptionally there is absolute flatness over the tumor.

It is often of assistance in determining that a tumor belongs to the pylorus to find dilatation of the stomach. An abnormal fulness of the epigastric and umbilical regions may then be observed, and through the abdominal walls, if thin, may be seen the peristaltic movements of the stomach. Other signs and symptoms aid in the diagnosis of dilatation of the stomach, and will be described in connection with this disease.

It is to be noted that what one takes to be the primary tumor of the stomach is not so very rarely a secondary cancerous mass in the stomach or in adjacent lymph-glands or in the peritoneum. Such nodules may also increase the apparent size of the original tumor. As has been pointed out by Rosenbach,⁵⁰ spasm of the muscular coat near a cancer or an ulcer of the stomach may produce a false tumor or enlarge a real tumor.

⁵⁰ Deutsche med. Wochenschr., 1882, p. 22.

The cancer, instead of appearing as a circumscribed tumor, may infiltrate diffusely the gastric walls, and so escape detection. When the greater part or the whole of the stomach is the seat of this diffuse cancerous infiltration, a sense of abnormal resistance may be appreciated by palpation in the epigastric region. In these cases the stomach is often much shrunken in size. The outlines of the thickened organ can sometimes be made out, but the physical signs do not suffice for the diagnosis of cancer.

With cancer of the cardia there is usually more or less atrophy of the stomach, which is manifested by sinking in of the epigastric region.

Sometimes the tumor eludes discovery on account of special obstacles to the physical examination of the abdomen, such as a thick layer of fat in the abdominal walls or a large quantity of ascitic fluid. Every aid in the physical examination of the abdomen should be resorted to. The patient should be examined while lying on his back with the utmost possible relaxation of the abdominal walls. If necessary, he should also be examined while standing or in the knee-elbow position. Sometimes a deep inspiration will force down a previously concealed tumor. The emptying of a dilated stomach by means of a stomach-tube will sometimes bring to prominence a gastric tumor.

The inflation of the stomach by the development in it of carbonic acid gas may render valuable assistance in the diagnosis of tumors of this organ and of surrounding parts. This method has been recommended by W. Ph. H. Wagner among others, and especially by Rosenbach.⁵¹ From 20 to 30 grains of bicarbonate of soda and from 15 to 20 grains of tartaric acid may be introduced into the stomach. The soda, dissolved in lukewarm water, may be given first and followed by the acid in solution, or, better, the mixed powders may be swallowed in the dry state and followed by a tumblerful of water. Some persons require a larger quantity of the powder in order to inflate the stomach. Occasionally the introduction of the effervescing powder fails to produce any appreciable distension of the stomach. This negative result may be due to the escape of the gas into the intestine in consequence of incontinence of the pylorus—a condition which Ebstein⁵² has observed and described especially in connection with pyloric cancer. When this pyloric insufficiency exists the resulting tympanitic distension of the intestine is a hindrance to palpation of tumors of the stomach. Failure to secure distension of the stomach is not always due to this cause. It may be necessary to make repeated trials of the effervescing mixture. It is well to have a stomach-tube at hand to evacuate the gas if this should cause much distress.

⁵¹ W. Ph. H. Wagner, Ueber die Percussion des Magens nach Auftreibung mit Kohlensäure, Marburg, 1869; O. Rosenbach, Deutsche med. Wochenschr., 1882, p. 22.

⁵² W. Ebstein, Volkmann's Samml. klin. Vorträge, No. 155.

In some respects simpler and more easily controlled is the method of distending the stomach by injecting air into it through a stomach-tube, as recommended by Runeberg.⁵³ For this purpose the balloon of a Richardson's spray apparatus may be attached to a soft-rubber stomach-tube. In this way the desired quantity of air can be introduced and at any time allowed to escape through the tube.

⁵³ J. W. Runeberg, Deutsches Arch. f. kl. Med., Bd. 34, p. 460, 1884.

When the stomach has been inflated the contours of tumors of the pylorus often become surprisingly distinct in consequence of the changes in the position and the shape of the stomach. When the tumor is fixed by adhesions, it may be possible to follow the contours of the stomach into those of the tumor. False tumors produced by spasm of the muscular walls of the stomach may be made to disappear by this distension of the organ. This procedure enables one to distinguish between tumors behind and those in front of the stomach, as the former become indistinct or disappear when the stomach is inflated. By bringing out the contours of the stomach the relations of the tumor to surrounding organs may be rendered for the first time clear. Assistance in diagnosis may also be afforded by distension of the colon with water or with gas or with air, per rectum, in order to determine the course of the colon and its relations to abdominal tumors (Mader, Ziemssen, Runeberg). A manifest contraindication to distension of the stomach or of the colon with gas exists if there is a suspicion that the coats of these parts are so thinned by ulceration that they might rupture from the distending force of the gas. There have been no cases recorded where such an accident has happened.

Only in exceptional cases are the bowels regular throughout the course of gastric cancer. Constipation is the rule, and not infrequently there is obstinate constipation. This is to be expected when the patient eats little and vomits a great deal, or when there is stenosis of the pylorus. In cancer, as in many other diseases of the stomach, the peristaltic movements of the intestine are inclined to be sluggish.

Occasional diarrhoea is also common in gastric cancer, being present, according to Tripier,⁵⁴ at some period or other in over one-half the cases. Constipation often gives place to diarrhoea during the last months or during the last days of life. In other periods of the disease diarrhoea not infrequently alternates with constipation. In rare cases diarrhoea is an early symptom, and it may be present exceptionally throughout the greater part of the disease. The irritation of undigested food sometimes explains the diarrhoea. When diarrhoea is persistent there probably exists catarrhal inflammation of the large intestine, or in some instances there may be diphtheritic and ulcerative inflammation of the colon, causing dysenteric symptoms during the last stages of cancer of the stomach.

⁵⁴ "Étude clinique sur la Diarrhée dans le Cancer de l'Estomac," Lyon Méd., 1881, Nos. 40, 41, 42.

Black stools containing altered blood occur for some days after profuse gastric hemorrhage. It is important to examine the stools for blood, as bleeding may occur from cancer of the stomach without any vomiting of blood.

There is no change in the urine characteristic of gastric cancer. Deposits of urates are not uncommon. If there be profuse vomiting or frequent washing out of the stomach, the urine often becomes alkaline from fixed alkali.⁵⁵ The amount of urea is diminished in consequence of the slight activity of the nutritive processes of the body. Rommelaere attaches unmerited diagnostic importance to this diminution of urea. A similar diminution of urea occurs in other like states of depressed nutrition.

⁵⁵ According to Quincke, when the acid in the stomach is not hydrochloric acid, but organic acid resulting from fermentation, then vomiting and washing out the stomach do not reduce the acidity of the urine (Zeitschrift f. klin. Med., Bd. 7, Suppl. Heft, p. 25).

Albuminuria does not belong to the history of gastric cancer, although a small quantity of albumen may be present in the urine as in other anæmic and cachectic conditions. A larger quantity of albumen may be due to parenchymatous and fatty degeneration of the kidney or to chronic diffuse nephritis, which are infrequent but recognized complications of gastric cancer. There is often an excess of indican in the urine, to which, however, no diagnostic significance can be attached.

The urine in gastric cancer sometimes contains an excess of aceton, or at least of some substance which yields aceton upon the application of various tests. This so-called acetonuria is present without any symptoms referable to it, so far as we know. Allied to this so-called acetonuria is that condition of the urine in which it is colored burgundy-red upon the addition of ferric chloride in solution (Gerhardt's reaction). It is not positively known what substance imparts this last reaction to the urine. V. Jaksch, who has studied the subject industriously, believes that the red coloring substance is diacetic acid, and he proposes to call the condition diaceturia. Fresh urine, which shows in a marked degree Gerhardt's reaction, often has a peculiar aromatic, fruity odor, as has also the expired air. Gerhardt's reaction has been studied mostly in diabetic urine, but it occurs sometimes in cases of gastric cancer and in a variety of diseases. This so-called diaceturia may be associated with a peculiar form of coma, but it is oftener observed without any symptoms referable to it⁵⁶ (see page 555).

⁵⁶ The various tests for aceton in the urine are not altogether satisfactory. They are to be found in an article by Von Jaksch in the Zeitschrift f. klin. Med., Bd. viii. p. 115. For English readers a good abstract of an article by Penzoldt on these tests and on acetonæmia in general is to be found in The Medical News of Philadelphia, Aug. 9, 1884, p. 162, but this does not consider the corrections and additions to be found in V. Jaksch's article cited above. Acetonuria has been observed especially in diabetes mellitus, fevers, carcinoma, and dyspepsia.

The substance which produces Gerhardt's reaction is to be distinguished from other substances which may be present in the urine and give a red color with ferric chloride—first, by the fact that boiling the urine in a test-tube for five or six minutes destroys the first-named substance, or causes the red color to disappear in case this has been produced by ferric chloride; and, secondly, by the fact that ether extracts the substance from acidified urine, and that the red color produced in the ether extract by ferric chloride (it may be necessary to first neutralize the acid) fades away in the course of a few days (V. Jaksch, Zeitschrift f. Heilkunde, Bd. iii. p. 17). Urines which respond to Gerhardt's reaction in a marked degree yield aceton on distillation, but aceton or an aceton-yielding substance may be present in considerable quantity without response of the urine to Gerhardt's test.

Disorders of nutrition embrace an important group of symptoms, such as loss of flesh and strength, impoverished blood, and cachectic color of the skin. Emaciation and debility are sometimes the first symptoms of gastric cancer to attract attention, and often the first symptoms to arouse anxiety. More frequently these symptoms of disordered nutrition first appear after dyspeptic ailments or pain have existed for several weeks or months. It may aid in the diagnosis of gastric cancer to weigh the patient from time to time, as carcinoma is generally attended by progressive loss of weight.

The patient frequently becomes morose and depressed in spirits. His strength fails, sometimes disproportionately to the loss of flesh. There is no disease in which emaciation becomes more extreme than in cases of gastric cancer.

In many cases profound anæmia develops, and sometimes in such a degree that this symptom cannot be regarded always as simply co-ordinate with the other disorders of nutrition, but is to be regarded rather as an evidence of some special disturbance of the blood-forming organs. The blood may present the same changes as are observed in pernicious anæmia, such as extreme reduction in the number of red blood-corpuscles (to one million or even half that number in a cubic millimeter) and manifold deformed shapes of the corpuscles (poikilocytosis). In extreme cases the proportion of hæmoglobin in the blood may be reduced to 50 or 60 per cent. of the normal quantity.⁵⁷ There is occasionally a moderate increase in the number of white blood-corpuscles. In one case of gastric cancer I observed a leucocytosis in which there was one white to twenty red blood-corpuscles without enlargement of the spleen.⁵⁸

⁵⁷ The granular disintegrating corpuscles (Zerfallskörperchen of Riess) may also be found in the blood in considerable number. Leichtenstern has observed that toward the end of life the relative proportion of hæmoglobin in the blood may be increased, sometimes rapidly, and may even exceed the normal limit. This is due to concentration of the blood in consequence of the loss of water. In such cases the tissues appear abnormally dry and the blood thick and tarry at the autopsy (Ziemssen's Handb. d. spec. Path. u. Therap., Bd. viii. 1te Hälfte, p. 344).

It seems to me proper to distinguish two kinds of anæmia in gastric cancer—a simple anæmia, which is present in the majority of cases, and can be explained by the development of the cancer and the disturbance of the gastric functions; and a pernicious anæmia, which is present only in exceptional cases, and has the typical symptoms of progressive pernicious anæmia.

⁵⁸ In a case of large medullary cancer of the stomach reported by H. Mayer there was one white to fifty red blood-corpuscles. The spleen was not enlarged (Bayer, Aerztl. Intelligenzblatt, 1870, No. 21). A similar case is related by Lebert, in which, however, the spleen was enlarged (op. cit., p. 481).

To the pallor of anæmia is added often a faded yellowish tint of the skin which is considered characteristic of the cancerous cachexia. At the same time, the skin is frequently dry and harsh, and may present brownish spots (chloasma cachecticorum). The pallid lips, the pale greenish-yellow color of the face, the furrowed lines, and the pinched and despondent expression make up a characteristic physiognomy, which, however, is neither peculiar to gastric cancer nor present in all cases of the disease. There is no cachectic appearance which is pathognomonic of cancer; and in this connection it is well to note that there are cases of gastric ulcer, and particularly of non-cancerous stenosis of the pylorus, in which all of the symptoms described as peculiar to the cancerous cachexia are met with. Nevertheless, the weight of these symptoms in the diagnosis of gastric cancer should not be underestimated. There is no disease in which profound cachectic symptoms so frequently and so rapidly develop as in gastric cancer.

The profound nutritive disturbances of gastric cancer are referable partly to the cancer as such, and partly to the impairment of the functions of the stomach. It is impossible to separate the effects of these two sets of causes, and distinguish, as some have done, a cachexia of cancer and a cachexia of inanition. It is the combination of these causes which renders the cachexia of cancer of the stomach so common, so rapid in its development, and so profound as compared with that of cancer in other situations. The relation of cancer in general to cachexia need not here be discussed, save to say that there is the best ground for believing that the cachexia is directly dependent upon the growth and metamorphoses of the primary cancer and its metastases, and that there is not reason to assume any dyscrasia antedating the cancerous formation.

While the failure of the general health and the gastric symptoms in general develop side by side, it is especially significant of gastric cancer when the symptoms of impaired nutrition are more pronounced than can be explained by the local gastric disturbance. When, however, as sometimes happens, gastric symptoms are absent or no more than can be explained by anæmia and marasmus, then in the absence of tumor a positive diagnosis is impossible. Such cases of gastric cancer during life often pass for essential or pernicious anæmia. Otherwise, unexplained symptoms of anæmia with emaciation and debility, particularly in elderly people, should lead to a careful search for gastric cancer.

Finally, it is necessary to add that there are exceptional cases of gastric cancer in which there is no emaciation, and in which the general health appears to be astonishingly well preserved. In most of these cases death occurs either from some accident of the disease or from some complication.

Slight or moderate oedema about the ankles is a common symptom during the cachectic stage of gastric cancer. This oedema is due to hydræmia. This cachectic dropsy in rare cases becomes excessive and leads to anasarca, with serous effusion in the peritoneal, pleural, and pericardial sacs. Such cases are liable to be mistaken for heart disease, particularly as a hæmic murmur often coexists, or for Bright's disease. Ascites may be the result not only of hydræmia, but also of cancerous peritonitis or of pressure on the portal vein by cancer. Many cases of gastric cancer associated with ascites have been falsely diagnosed as cirrhosis of the liver, and sometimes the distinction is extremely difficult or impossible.

During the greater part of the disease the pulse is usually normal; toward the end it is not infrequently rapid, small, and compressible. In consequence of weakness and anæmia any exertion may suffice to increase the frequency of the pulse, and may induce palpitation of the heart and syncope.

As might be expected as the result of anæmia, hæmic murmurs in the heart and blood-vessels are not rare in gastric cancer.

Epigastric pulsation is often very prominent in cases of gastric cancer, as it may be in various other conditions. This pulsation is sometimes of a paroxysmal nature.

Venous thrombosis is not a rare complication in the last stages of gastric cancer. It is most common in the femoral and saphenous veins, and is rapidly followed by painful oedematous swelling of the affected extremity. Thrombosis of the subclavian and axillary veins is much less frequent. When it occurs there are the same symptoms of phlegmasia alba dolens in the upper extremity as have been mentioned for the lower. Lebert has recorded a case of thrombosis of the right external jugular vein.⁵⁹ The thrombosis is the result of marasmus, and therefore may occur in other gastric diseases besides gastric cancer, so that this symptom has not all the diagnostic importance for gastric cancer claimed by Trousseau. Being an evidence of great weakness of the circulation, marantic thrombosis in cancer of the stomach is of grave prognostic import.

⁵⁹ Op. cit., p. 394.

The temperature is often normal throughout the course of gastric cancer. Febrile attacks, however, are not uncommon in this disease. Elevation of temperature may occur without any complication to explain it. During the second half of the disease there may be either irregular febrile attacks or a more continuous fever, which is, however, usually of a light grade, the temperature not generally exceeding 102°. Slight chills may be experienced. Lebert describes a light and a hectic carcinomatous fever.

There may be subnormal temperature with collapse during the last days of life, and in general anæmia and inactivity of nutritive processes tend to produce a low temperature.

Dyspnoea on slight exertion may be present in gastric cancer as a result of anæmia or of fatty heart. In a few cases of gastric cancer have been observed symptoms pointing to a reflex vagus neurosis, such as paroxysms of dyspnoea, oppression in the chest, and palpitation of the heart, but these symptoms are less common in gastric cancer than in some other diseases of the stomach. Watson⁶⁰ relates a case of gastric cancer in which increasing dyspnoea and palpitation were such prominent symptoms that he was led to diagnose fatty heart with portal congestion as the sole trouble. At the autopsy the heart and lungs were found healthy, but there was extensive cancer of the greater curvature of the stomach. He subsequently ascertained that there had been symptoms pointing to gastric disease.

⁶⁰ Sir T. Watson, Lectures on the Principles and Practice of Physic, vol. ii. p. 471, Philada., 1872.

The various complications of gastric cancer which affect the respiratory organs will be considered later.

Depression of spirits, lack of energy, headache, neuralgia, sleeplessness, and vertigo are functional nervous disturbances which are often the result of disordered digestion from whatever cause, and are therefore not uncommon in gastric cancer. The theory that these symptoms are due to the absorption of noxious substances produced in the stomach and intestine by abnormal digestive processes is plausible,⁶¹ and more intelligible than reference to some undefined sympathy between the digestive organs and the nervous system.

⁶¹ This theory is elaborated by Senator ("Ueber Selbstinfection durch abnorme Zersetzungsvorgänge, etc.," Zeitschrift f. klin. Med., Bd. 7, p. 235).

The intelligence is generally not impaired in the course of gastric cancer.

Considerable interest belongs to coma as a symptom of cancer of the stomach, and more particularly to the occurrence of coma with the peculiar characters which have been described by Kussmaul as distinguishing diabetic coma.⁶² The most distinctive feature in Kussmaul's group of symptoms is the accompaniment of the coma by a peculiar dyspnoea in which, without evidence of disease of the lungs or air-passages, the respirations are strong and deep and often attended with a groaning sound in expiration. The breathing is either normal in frequency or oftener moderately increased. The pulse is usually small and frequent. The temperature is not much elevated, and sometimes is much below the normal. Sometimes the coma is preceded by a period of excitement, with restlessness, and perhaps with screaming. Gerhardt's reaction in the urine may or may not be present. When it is present in a marked degree there is often an aromatic, chloroform-like odor to the breath and to the fresh urine. The patient may come out of the coma, but in the vast majority of cases the coma terminates fatally.

⁶² Deutsches Arch. f. klin. Med., Bd. 14, p. 1.

It is now known that this dyspnoeic coma is not confined to diabetes mellitus, but that it occurs also in gastric cancer and in various other diseases.⁶³ Its occurrence in gastric cancer is rare. In this disease it does not usually appear until anæmia is far advanced, but it may occur in cases of cancer in which the patient's general health and nutrition are still fairly good. I recently made the post-mortem examination of an elderly man, fairly well nourished, who was found in the streets comatose and brought in this condition to Bellevue Hospital, where he died in about twelve hours. While in the hospital his breathing was increased in frequency, forcible, and deep. His temperature was normal. The urine contained a small quantity of albumen, but no sugar. No previous history could be obtained. Uræmic coma was suspected. At the autopsy was found a large, soft, ulcerated cancer of the lesser curvature and posterior wall of the stomach near the pylorus. The kidneys, brain, heart, and other organs were essentially healthy.

⁶³ Von Jaksch was the first to describe this form of coma in cancer of the stomach (Wien. med. Wochenschr., 1883, pp. 473, 512). He adopted the term coma carcinomatosum, and more recently coma diaceticum. L. Riess has reported seventeen cases of this coma occurring in a variety of diseases, such as pernicious anæmia, gastric cancer, gastric ulcer, tuberculosis, which all had in common profound anæmia. He proposes the term dyspnoeic coma (Zeitschrift f. klin. Med., Bd. 7, Suppl. Heft, p. 34, 1884). Senator has described two cases of gastric cancer with this coma. He uses the terms dyscrasic coma and Kussmaul's group of symptoms (ibid., Bd. 7, p. 235). In the cases described by Litten under the name coma dyspepticum, dyspnoea was absent, but Gerhardt's reaction in the urine was present. In Litten's cases structural disease of the stomach was not supposed to be present. The patients recovered from the coma (ibid., Suppl. Heft, p. 81).

We possess no satisfactory explanation of this form of coma. In diabetes it is considered to be due to the presence in the blood of some intoxicating agent. For a time this agent was thought to be aceton; it is now believed by Von Jaksch to be diacetic acid. Much stress has been laid upon the aromatic, fruity odor of the breath and of the fresh urine, and upon the presence of some substance in the urine which imparts to it a burgundy-red color upon the addition of liquor ferri chloridi (Gerhardt's reaction. See changes in the urine, page 551). Although the whole aceton question is at present in a very confused state, there is no proof that aceton or its allies possesses the toxic properties assumed by this theory;⁶⁴ and it is certain that dyspnoeic coma may occur in diabetes and in other diseases without the presence of Gerhardt's reaction in the urine. It is also true that this reaction often occurs without any clinical symptoms referable to it. Riess and Senator believe that in non-diabetic cases anæmia is the most important factor in the production of this coma.⁶⁵

⁶⁴ Frerichs, Zeitschrift f. klin. Med., Bd. 6, p. 3.

⁶⁵ Riess refers the coma to the anæmia as such, whereas Senator thinks that, in consequence of the depraved nutrition of the body resulting from the anæmia, some toxic substance is developed which enters the circulation.

Coma, probably belonging to this same variety, may occur in gastric cancer without the peculiar dyspnoea which has been described. There is reason to believe that this dyspnoea is not a necessary symptom of the so-called diabetic coma.

Chronic Bright's disease terminating with uræmic coma is an occasional but not frequent complication of gastric cancer.

Coma and other cerebral symptoms may be produced by secondary cancerous tumors in the brain.

Stupor deepening into coma may develop during the often-prolonged death-agony of gastric cancer.

The distribution, origin, and frequency of cancerous growths secondary to gastric cancer are most conveniently considered under Pathological Anatomy. Symptoms referable to certain localizations of these secondary cancerous deposits, however, are so common, and so interwoven with the clinical history of cancer of the stomach, that it is desirable to consider some of these symptoms in the present connection.

Cancer of the liver is the most important of these secondary cancerous growths. It is estimated to be present in nearly one-third of the cases of gastric cancer, but by no means in all these cases does it produce symptoms. As a rule, the earlier hepatic cancer forms in the course of gastric cancer the more likely is it to be attended by symptoms. The most important symptoms of secondary cancer of the liver are enlargement of the liver, peritoneal exudation, and persistent icterus. When nodular growths can be felt in the free border or surface of the liver, the diagnosis is generally easily established. Sometimes the liver remains of normal size or is even contracted, and then the diagnosis is difficult or impossible. Ascites or exudative peritonitis is present in about one-half of the cases of cancer of the liver. Jaundice is less frequently present. It is only persistent jaundice which aids in the diagnosis of hepatic cancer.

The various combinations of gastric cancer with secondary hepatic cancer may be clinically grouped as follows:

1. Symptoms of gastric cancer with latent hepatic cancer.

2. Symptoms of gastric cancer followed by symptoms of hepatic cancer.

3. Symptoms both of gastric cancer and of hepatic cancer present when the case comes under observation.

4. Symptoms of hepatic cancer with latent gastric cancer.

5. Symptoms of hepatic cancer followed by symptoms of gastric cancer.

6. Both hepatic and gastric cancer latent. Symptoms of anæmia and marasmus, or of chronic exudative peritonitis, or of chronic pleurisy.

From this grouping it is evident that the existence of secondary hepatic cancer may aid in the diagnosis of cancer of the stomach, or may mislead, or may be without influence. The greatest assistance in diagnosis is rendered when the physical signs and the symptoms of hepatic cancer develop some time after the appearance of gastric symptoms which may previously have been equivocal. Much more difficult to diagnosticate are the cases of hepatic cancer accompanied or followed by gastric symptoms, inasmuch as cancer of the liver, whether primary or secondary, may be attended with marked disturbance of the gastric functions, including hæmatemesis. In these cases, unless a tumor of the stomach can be discovered, a positive diagnosis of gastric cancer is impossible. In view of the infrequency of primary cancer of the liver, however, there will be in many of these cases a strong probability in favor of primary cancer of the stomach. When it is remembered that over one-third of the cancers of the liver are secondary to cancer of the stomach, it is evident that in cases which appear to be primary hepatic cancer very careful attention should be given to the exploration of the stomach. But even then diagnostic errors will often be unavoidable.

Cancer of the peritoneum secondary to cancer of the stomach may produce no symptoms, and so pass unrecognized. The diagnosis of peritoneal cancer is readily made when, after the recognition of gastric cancer, secondary cancerous nodules in the peritoneum can be felt through the abdominal walls or through the vagina. There are cases of gastric cancer in which the symptoms are all referable to secondary cancer of the peritoneum. Cancer of the peritoneum is usually attended with fluid exudation in the peritoneal cavity. The chemical and the microscopical examination of this fluid withdrawn by paracentesis may aid in the diagnosis of cancerous peritonitis. Whereas in dropsical accumulations in the peritoneal cavity the quantity of albumen in the fluid is usually less than 2½ per cent., in cancerous peritonitis there is usually from 3 to 4 per cent. of albumen, the percentage rarely falling as low as 2½ per cent., but sometimes being as high as from 5 to 6 per cent. The percentage of albumen in ordinary peritonitis is usually over 4.⁶⁶ Clumps of cancer-cells are sometimes to be found by microscopical examination of the fluid. These cells are large, epithelioid in shape, and often contain vacuoles and fatty granules. It is only when these cells are arranged in clumps or as so-called budding cells, and when they are present in abundance, that they are diagnostic. They are to be sought especially in fibrinous coagula. They are present only when the cancerous alveoli actually communicate with the peritoneal cavity.⁶⁷ The development of cancerous nodules in the margins of an opening made in the abdominal walls by a trocar is also evidence of cancerous disease of the peritoneum. The same thickening and retraction of the mesentery and omentum may occur in cancerous as in tuberculous peritonitis. In both the exudation is often hemorrhagic.

⁶⁶ The conditions under which the estimation of the quantity of albumen in the peritoneal exudation may prove of diagnostic aid are fully considered by Runeberg (Deutsches Arch. f. klin. Med., Bd. 34, p. 1). Here also are given methods for making this analysis for clinical purposes.

⁶⁷ The literature on this subject is as follows: Foulis, Brit. Med. Journ., July 20, Nov. 2, 1878; Thornton, ibid., Sept. 7, 1878; Quincke, Deutsches Arch. f. klin. Med., Bd. 30, p. 580; Ehrlich, Charité Annalen, vii. p. 226; Brieger, ibid., viii.

Importance has been attached to enlargement of the supraclavicular lymphatic glands in the diagnosis of cancer of the stomach, but there are so many causes of enlargement of these glands that not much significance can be attached to this symptom, which, moreover, is absent in most cases. Still, under certain circumstances this glandular enlargement may aid in the diagnosis. The same remarks apply to enlargement of the inguinal glands, which is a common occurrence in case cancer involves the peritoneum. One must not mistake abnormal prominence of the lymphatic glands in consequence of emaciation for actual enlargement.

Gastric cancer much less frequently than gastric ulcer causes perforation of the stomach. Of 507 cases of gastric cancer collected by Brinton, perforation into the general peritoneal cavity occurred in 17 (3¹/3 per cent.).⁶⁸ In two cases of gastric cancer reported by Ellis perforative peritonitis was preceded by symptoms supposed to be only those of ordinary dyspepsia, hemorrhage and vomiting being absent.⁶⁹ Various fistulous communications like those described under gastric ulcer may be the result of perforation of gastric cancer, but with the exception of gastro-colic fistula they are much more frequently produced by ulcer than by cancer. In 160 cases of gastric cancer collected by Dittrich, gastro-colic fistula existed in 6 (3¾ per cent.).⁷⁰ In 507 cases collected by Brinton this fistula existed in 11 (2.17 per cent.). In Lange's 210 cases gastro-colic fistula existed in 8 (3.8 per cent.). Of 33 cases of gastro-colic fistula collected by Murchison, 21 were caused by cancerous ulceration.⁷¹ The symptoms characteristic of fistulous communication between the stomach and the colon are the vomiting of fecal matter and the passage of undigested food by the stools. These symptoms are not present in all cases, so that a diagnosis is not always possible. Fecal vomiting is influenced by the size of the opening between the stomach and the colon. With great obstruction at the pylorus, fecal vomiting, as might be expected, is absent or infrequent, while the passage of undigested food by the bowels is common. Under these circumstances vomiting is sometimes relieved after the establishment of the fistula. Aid may be afforded in the diagnosis of gastro-colic fistula by the introduction into the rectum or into the stomach of colored or other easily recognizable substances, and determining their presence in the vomit or in the stools in consequence of their escape by the unnatural outlet. V. Ziemssen has determined in a case of gastro-colic fistula due to cancer the escape into the stomach of carbonic acid gas artificially generated in the rectum, with failure to obtain distension of the colon.⁷² A number of instances of gastro-cutaneous fistula due to gastric cancer have been recorded, but this form of fistula is much less common than gastro-colic fistula, and much less frequently the result of cancer than of ulcer of the stomach. Subcutaneous emphysema may precede the formation of the fistula. Other gastric fistulous communications resulting from cancer, such as with the pleura, the lungs, the small intestine, are too infrequent to merit consideration under the symptomatology of the disease.

⁶⁸ Loc. cit. Lange (op. cit.) records in 210 cases of gastric cancer 12 perforations into the peritoneal cavity (5.7 per cent.).

⁶⁹ Extr. fr. the Rec. of the Boston Soc. for Med. Improvement, vol. iii. p. 116, and vol. iv. p. 109.

⁷⁰ Prager Vierteljahrsch., vol. xvii.

⁷¹ Edinb. Med. Journ., vol. iii. p. 4, 1857.

⁷² Deutsches Arch. f. kl. Med., Bd. 33, p. 237. He recommends for extreme distension of the colon in an adult the introduction, by means of a tube passed up the rectum, of a solution of about 5 drachms of sodii bicarb. and 4½ drachms of tartaric acid—injected not all at once, but in three or four doses at intervals of a few minutes, the tube being cleaned in the intervals by the injection of three ounces of water, so as to avoid generation of gas in the tube. The generation of a smaller quantity of gas would suffice for the purpose here in view.

As a rule, patients with gastric cancer die from gradual exhaustion. In a condition of extreme emaciation and feebleness the patient sinks into a state of collapse, accompanied often with stupor, sometimes with mild delirium. The death-agony is prolonged frequently from twelve to twenty-four hours, and sometimes even longer. On the other hand, death may occur somewhat suddenly in the last stages of gastric cancer, and without satisfactory explanation.

Death from copious gastric hemorrhage does not occur probably in more than 1 per cent. of the cases of cancer of the stomach.

In the rare cases of death from perforation of the stomach the patient is sometimes so exhausted at the time of perforation that the occurrence of this accident remains unrecognized in the absence of any complaint of characteristic symptoms.

The coma which sometimes leads to the fatal termination of gastric cancer has already been sufficiently considered.

Finally, death may be the result of certain complications more or less dependent upon the cancer. Of these the most important are suppurative peritonitis and pulmonary complications, particularly oedema, terminal pneumonia, and embolism of the pulmonary artery.

DURATION.—It is evidently impossible to determine the exact duration of a cancer of the stomach. Doubtless in all cases there is a period of growth of the tumor before it produces symptoms, and the duration of this latent period can never be determined. When symptoms appear they are often at first so mild as to be readily overlooked, and so ambiguous that even if recognized they are not clearly referable to the cancer. Gastric symptoms may have preceded, perhaps for years, the development of the cancer, so as to lead to the assumption of a longer duration of the cancer than is really the case. Estimates, therefore, of the duration of gastric cancer can be only of limited value.

From 198 cases Brinton⁷³ estimates the average duration of gastric cancer as about twelve and a half months, the maximum duration as about thirty-six months, and the minimum as one month. From 36 cases Katzenellenbogen⁷⁴ estimates the average duration as eighteen months, the maximum as five years and five months, the minimum as one month. From 112 cases Lebert⁷⁵ makes the average duration fifteen months and the maximum four years. In 4 per cent. of the cases Lebert found the duration less than three months, in 62 per cent. between six and eighteen months, in 42 per cent. between six and twelve months, in 17 per cent. between three and six months, and in the same number of cases between eighteen months and four years.

⁷³ Loc. cit.

⁷⁴ Op. cit.

⁷⁵ Op. cit.

Estimates of several years' duration (such as nine years in the case of Napoleon) are to be received with scepticism. In these cases symptoms of gastralgia or of dyspepsia or of gastric ulcer have preceded the development of the cancer. It has already been mentioned that cancer may develop in a simple ulcer of the stomach.

Mathieu,⁷⁶ from an analysis of 27 cases of gastric cancer occurring under thirty-four years, found the average duration in early life to be only three months. In only 2 out of 19 cases did the duration exceed one year. Although this analysis is based upon too small a number of cases, there seems to be no doubt that gastric cancer pursues a more rapid course in early life than it does in old people.

⁷⁶ Du Cancer précoce de l'Estomac, Paris, 1884, p. 40.

COMPLICATIONS.—Some of the complications of gastric cancer have been mentioned under Symptomatology. Jaundice may appear in the course of gastric cancer from a variety of causes, such as catarrhal gastro-duodenitis, impaction of gall-stones in the common bile-duct, and pressure on the bile-duct by cancerous growths in the pancreas, in the portal lymphatic glands, or in the liver itself. Pylethrombosis, which is likely to be suppurative, is a rare complication. In a case of cancer of the anterior wall and greater curvature of the stomach reported by Wickham Legg⁷⁷ the symptoms seem to have been mostly referable to a complicating suppurative pylethrombosis. Simple and cancerous pylethromboses also occur. Other forms of peritonitis than the cancerous may complicate gastric cancer, such as suppurative, sero-fibrinous, and chronic proliferative peritonitis. Catarrhal enteritis, and particularly diphtheritic colitis, are not infrequent complications, especially in the later stages of the disease. Chronic diffuse nephritis, both in the form of the large and of the small kidney, is a rare complication of cancer of the stomach. Hydrothorax, sero-fibrinous pleurisy, and emphysema may develop either with or without cancerous invasion of the pleura. Pericarditis is much less common; it is most likely to occur with cancer of the cardia. Pyo-pneumothorax, abscess, and gangrene of the lung may result from perforation of the pleura or of the lung by gastric cancer. Oedema of the lungs, splenization, and pneumonia, involving usually the lower lobes, are common in the last days of gastric cancer. Emboli derived from venous thrombi are sometimes carried into the pulmonary artery or its branches. Although much has been written as to the exclusion of tuberculosis by cancer, no such law exists. Both old and fresh tubercles have been repeatedly observed in cases of gastric cancer. Reference has already been made to the frequent development of aphthæ in the mouth, pharynx, and oesophagus in the final stage of gastric cancer. Fatty degeneration of the heart may develop in gastric cancer as in other anæmic states. Phlegmasia alba dolens has already been mentioned. It is not probable that insanity is to be regarded as more than an accidental complication of gastric cancer; still, it has been noticed in several cases—for instance, of Dittrich's 160 cases, 5 patients were insane, 2 with violent mania. Amyloid degeneration has been present in some cases. Purpura hæmorrhagica has been present in a few instances in the later stages (cachectic purpura). Chronic catarrhal gastritis and dilatation of the stomach are less complications than a part of the disease. The relation of cancer to simple ulcer of the stomach has already been considered. The various secondary cancerous deposits are most conveniently considered under the Morbid Anatomy. It is to be remarked that many of the complications of gastric cancer—as, for instance, pneumonia and peritonitis—may have a very obscure clinical history, as they often occur when the patient is greatly prostrated.

⁷⁷ St. Bartholomew's Hosp. Rep., vol. x. p. 236.

MORBID ANATOMY.—The following table gives the situation of the tumor in 1300 cases of cancer of the stomach:⁷⁸

From this table it appears that three-fifths of all gastric cancers occupy the pyloric region, but it is not to be understood that in all of these cases the pylorus itself is involved. In four-fifths of the cases the comparatively small segment of the stomach represented by the cardia, the lesser curvature, and the pyloric region is the part affected by gastric cancer. The lesser curvature and the anterior and the posterior walls are involved more frequently than appears from the table, inasmuch as many cancers assigned to the pyloric region extend to these parts. The fundus is the least frequent seat of cancer. In the cases classified as involving the greater part of the stomach the fundus often escapes.

⁷⁸ These cases are collected from the following sources: Lebert, op. cit.; Prague statistics of Dittrich, Engel, Wrany, and Eppinger, loc. cit.; Habershon, op. cit.; Katzenellenbogen, op. cit.; and Gussenbauer and V. Winiwarter, loc. cit. Gussenbauer and V. Winiwarter assign to the class of cancers involving the whole stomach all cases which they found designated simply as carcinoma ventriculi without further description. This produces in their statistics an excessive number of cancers under this class. I have preferred, therefore, to estimate in their collection of cases the number of cancers involving the whole stomach, according to the percentage for this class obtained from the other authors above cited.

As was shown by Rokitansky, it is the exception for cancer of the pylorus to extend into the duodenum, whereas cancer of the cardia usually invades for a certain distance the oesophagus.

The varieties of carcinoma which develop primarily in the stomach are scirrhous, medullary, colloid, and cylindrical epithelial carcinoma.⁷⁹ The distinction between scirrhous and medullary cancer is based upon the difference in consistence, the former being hard and the latter soft. Cylindrical-celled epithelioma cannot be recognized as such by the naked eye. It presents usually the gross appearances of medullary cancer. Soft cancer (including both cylindrical-celled epithelioma and medullary carcinoma) is the most frequent form of gastric cancer. Next in frequency is scirrhous cancer, and then comes colloid cancer, which, although not rare, is much less frequent than the other varieties.

⁷⁹ I have not been able to find an authentic instance of primary melanotic cancer of the stomach, although this form is included by most authors in the list of primary gastric cancers. It is known that most cases formerly described as melanotic cancers are melanotic sarcomata, which originate usually in the skin or the eye and are accompanied frequently with abundant metastases. Secondary melanotic tumors have been several times found in the stomach. They were present in 7 out of 50 cases of melanotic cancer (or sarcoma) analyzed by Eiselt, although out of 104 cases not a single primary melanotic cancer occurred in the stomach (Prager Viertaljahrschr., vol. lxxvi. p. 54). The list of secondary melanotic sarcomata of the stomach might be still further increased. Of course gastric cancers colored by pigment from old blood-extravasations should not be confounded with melanotic tumors.

As all degrees of combination and of transition exist between the different forms of cancer, and as a large number of cancers of the stomach are of a medium consistence and would be classified by some observers as scirrhous and by others as medullary, statistics as to the relative frequency of the different varieties have very little value. Moreover, in most statistics upon this point there is no evidence that simple fibrous growths have not been confounded with scirrhous cancer, and as a rule little or no account is taken of cylindrical-celled epithelioma, which is a common form of gastric cancer—according to Cornil and Ranvier, the most common.⁸⁰

⁸⁰ For any who may be interested in such statistics I have collected 1221 cases of gastric cancer, of which 791 (64.8 per cent.) were medullary, 399 (32.7 per cent.) scirrhous, and 31 (2.5 per cent.) colloid. 22 cases described as epithelial have been included with the medullary; 29 cases described as fibro-medullary, and 1 as fasciculated, have been included with the scirrhous. The cases are from the previously-cited statistics of Lebert, Dittrich, Wrany, Eppinger, Gussenbauer, and V. Winiwarter, and from Fenger (Virchow u. Hirsch's Jahresbericht, 1874, Bd. i. p. 312).

Cancer of the stomach may grow in the form of a more or less complete ring around the circumference of the stomach, or as a circumscribed tumor projecting into the cavity of the stomach, or as a diffuse infiltration of the walls of the stomach. The annular form of growth is observed most frequently in the pyloric region. Cancerous tumors which project into the interior of the stomach are sometimes broad and flattened, sometimes fungoid in shape, but most frequently they appear as round or oval, more rarely irregular, crater-like ulcers, with thickened, prominent walls and ragged floor. The free surface of the tumor presents sometimes a cauliflower-like or dendritic appearance, which characterizes the so-called villous cancer. Diffuse cancerous infiltration is seated oftenest in the right half of the stomach, but it may occupy the cardiac region or even the entire stomach.

The relation of the cancerous growth to the coats of the stomach varies in different cases. The tumor usually begins in the mucous membrane and rapidly extends through the muscularis mucosæ into the submucous coat. In this lax connective-tissue coat the tumor spreads often more rapidly than in the mucous membrane, so that it may appear as if the cancer originated in the submucosa. The mucous membrane, however, is usually invaded, sooner or later, over the whole extent of the tumor. The dense muscular coat offers more resistance to the invasion of the tumor. Cancerous masses, however, penetrate along the connective-tissue septa between the muscular bundles, which often increase in number and size. In the muscular coat thus thickened can be seen the opaque white fibrous and cancerous septa enclosing the grayish, translucent bundles of smooth muscular tissue. Often, however, the whole muscular coat beneath the tumor is replaced by the cancerous growth, and can no longer be recognized. The serous and subserous connective tissue, like the submucous coat, offers a favorable soil for the growth of the tumor, which here appears usually in the form of large and small nodules projecting from the peritoneum. Adhesions now form between the stomach and surrounding parts, and opportunity is offered for the continuous growth of the cancer into these parts. In the manner described the tumor grows in all directions, sometimes more in depth, sometimes more laterally, sometimes more into the interior of the stomach.

Ulceration occurs in all forms of gastric cancer.⁸¹ The ulceration is caused either by fatty degeneration and molecular disintegration of the surface of the tumor or by the separation of sloughy masses. Doubtless the solvent action of the gastric juice aids in the process. The softer and the more rapid the growth of the cancer, the more extensive is likely to be the ulcer. Such ulcers are usually round or oval in shape, but their contours may be irregular from the coalescence of two or more ulcers or from serpiginous growth. The edges are usually high, soft in consistence, and often beset with polypoid excrescences. The floor is generally sloughy and soft, and often presents warty outgrowths. The edges and floor may, however, be hard and smooth. In the more slowly-growing scirrhous and colloid cancers the ulcers are more likely to be superficial. Partial cicatrization of cancerous ulcers may take place. The development of cicatricial tissue may destroy the cancerous elements to such an extent that only by careful microscopical examination can the distinction be made between cancer and simple ulcer or fibroid induration. The examination of secondary cancerous deposits in adjacent lymphatic glands or other parts becomes, then, an important aid in the diagnosis.

⁸¹ Ulceration was present in 60 per cent. of Lebert's cases, and in 66½ per cent. of Gussenbauer and V. Winiwarter's pyloric cancers.

Suppuration has been known to occur in gastric cancers, but it is extremely rare.

Each form of gastric cancer has certain peculiarities which require separate consideration.

Medullary carcinoma grows more rapidly than the other varieties of cancer. It forms usually soft masses, which project into the stomach and are prone to break down in the centre and develop into the crater-like ulcers already described. All of the coats of the stomach are rapidly invaded by the growth. The consistence of the tumor is soft, the color upon section whitish or reddish-gray, sometimes over a considerable extent hemorrhagic. Milky juice can be freely scraped from the cut surface of the tumor. The so-called villous cancer and the hæmatodes fungus are varieties of medullary carcinoma. Medullary carcinoma is more frequently accompanied by metastases than the other forms. In consequence of its tendency to deep ulceration medullary cancer is more liable to give rise to hemorrhage and to perforation than is scirrhous or colloid cancer. The continuous new formation of cancerous tissue in the floor of the ulcer and the formation of adhesions, however, greatly lessen the danger of perforation into the peritoneal cavity.

Histologically, medullary cancer is composed of a scanty stroma of connective tissue enclosing an abundance of cancerous alveoli filled with polyhedrical or cylindrical epithelial cells. The stroma is often richly infiltrated with lymphoid cells, and contains blood-vessels which often present irregular dilatations of their lumen.

Waldeyer describes with much detail, for this as for the other forms of gastric cancer, the origin of the tumor from the gastric tubules. According to his description, a group of gastric tubules, ten to twenty in number, sends prolongations downward into the submucous coat. These tubular prolongations are filled with proliferating epithelial cells, which make their way into the lymphatic spaces of the surrounding tissue and give origin to the cells in the cancerous alveoli. A small-celled infiltration of the surrounding connective tissue accompanies this growth of the tubules.