affect the secreting mechanism in these experiments. That there is an effect, these experiments illustrate very well. In experiment B the average of the hydrogen ion concentrations was pH-7.4 which is the same as that in experiment C. However, in experiment C 100 cc. of artery solution produced only 1.367 cc. of ureter solution as compared with 2.574 cc. of ureter solution for the same amount of artery solution in experiment B. In experiment C the smaller amount of secretion was correlated with a higher concentration of sodium chloride in the artery solution, namely 0.95 per cent as compared with 0.88 per cent in experiment B. A possible explanation of the lessened secretion in experiment C may be a force analogous to that which causes neutral salts to decrease quantitatively the swelling of colloids, as taught by Fischer. In these experiments it was not possible to take quantitative data to show whether secretion changes were accompanied by swelling changes, because of mechanical difficulties. That solution of protein took place has been noted, and it is known that in many colloids the solution process runs a course somewhat parallel to the swelling process, often accompanying it. Certainly the "kidneys" after the experiments were completed showed a typical picture of "cloudy swelling," but the change in size due to the blood pressure was such that minute changes in the size of colloids or their change back and forth from a precipitated form to a soluble form could not be measured under the conditions of the experiments. Furthermore, the process of cloudy swelling, with all the changes that go with it, evidently produces a change of elasticity in the kidney so that the same blood pressure will cause the kidney volume to increase more in the later stages than in the earlier (7).

The concentrations of the ureter solutions taken in grams of sodium chloride per cubic centimeter (see fig. 1) indicate that the secretion of sodium chloride is probably quantitatively not dependent on the secretion of water as, in the presence of a constant amount in the artery solution, the concentrations of some ureter solutions increased with the decrease in hydrogen ion concentration, whereas in others it decreased. In experiment C there is a consistent variation of the concentration of sodium chloride with that of the hydrogen ions, both increasing and decreasing together. This direct variation held true for only some of the other perfusions. In two cases (pH-7.4 in experiment C and pH-7.8 in experiment B) the repetition of a solution gave ureter solutions of practically the same concentration (number of grams of sodium chloride per cubic centimeter) but the relative amounts

excreted (amount of sodium chloride per 100 cc. of artery solution) were different, as pointed out before. If these results apply to the living kidney one would not expect to correlate the concentration of the urine directly with its acidity or with the hydrogen ion concentration of the blood, as the chart clearly shows. A third factor, namely, the colloid most affected at the hydrogen ion concentration in question, evidently must be taken into consideration.

The experiments on urea and on sodium chloride have shown that in the presence of a constant concentration of these substances in the artery solution, the amount secreted varied and the variation was closely related to the hydrogen ion concentration of the artery solution. They have also shown that although a chemical analysis may show the presence of a certain amount of urea or sodium chloride in the blood, yet the amount available for secretion may not be the same. Thus although an artery solution contained a constant amount of an aqueous salt solution, a variation of over 36 per cent excretion was noted in one experiment, the only variable factor being a change in hydrogen ion concentration from pH-7.2 to pH-7.8. While secretion of both urea and sodium chloride increases as the hydrogen ion concentration decreases, this relation does not hold true for all the colloids. in the body. Fischer (1) as stated before, finds that fibrin will hold sodium chloride, the more acid it is, while Reemelin finds that this does not hold for fibrin and urea. On this basis we can expect diminution in the chloride secretion ("salt or chloride retention") in the presence of an apparently good urea output, as, of the blood and kidney colloids, both tend to hold the chlorides, while of the two, only the kidney colloids tend to hold the urea. This condition is noted clinically in some conditions of chloride retention. Wolferth (8) notes that chloride retention in eclampsia may be marked while the retention of urea may be very slight.

The results of experiments on glucose, creatin, creatinine, uric acid and phenolsulphonephthalein will be reported later.

SUMMARY AND CONCLUSION

1. In isolated rabbit kidneys, perfused with salt solutions of constant and known composition, the secretion of sodium chloride and of water increases with the decrease in hydrogen ion concentration and decreases as the hydrogen ion concentration increases.

2. This relationship holds true for hydrogen ion concentrations within. physiological limits.

3. pH-6.6 probably represents the acid limit for function while the alkaline limit is pH-8.2 or higher. Variations beyond these are highly toxic to the kidney colloids and reduce secretion. The optimum hydrogen ion concentration lies between pH-7.2 and pH-7.8.

4. The amount of sodium chloride found in the blood on chemical analysis does not indicate the amount available for secretion. The amount available varies with the hydrogen ion concentration.

5. The effect of neutral salts (sodium chloride) in preventing the kidney colloids from taking up water and salt and secreting it can be shown quantitatively by varying the strength of the perfusing solution. 6. That the kidney colloids can be the source of albumin in the urine is shown by the fact that an artificial albuminuria, responding to the clinical tests, can be produced even though the artery solution contains no protein.

7. The amount of albumin washed out varies with the hydrogen ion concentration and follows the secretion of salt and water in which, under the conditions of the experiment, it becomes soluble.

8. There is evidence that some of the kidney proteins are more soluble in solutions of one hydrogen ion concentration than in another. This may account for the variations in kind of "albumin" found clinically in urine analyses, as well as the selective secretion of proteins.

BIBLIOGRAPHY

(1) FISCHER: Oedema and nephritis, 2nd. Ed., New York, 1915.

(2) REEMELIN AND ISAACS: This Journal, 1916, xlii, 163.

(3) REEMELIN: Lancet Clinic, 1916, cxv, 327.

(4) SOLLMANN: This Journal, 1905, xiii, 241.

(5) ISAACS: Anat. Rec., 1916, x, 206.

(6) TAYLOR, MILLER AND SWEET: Journ. Biol. Chem., 1917, xxix, 425.

(7) ISAACS: Anat. Rec., 1916, x, 517.

(8) WOLFERTH: Amer. Journ. Med. Sci., 1917, cliv, 84.

CONTRIBUTIONS TO THE PHYSIOLOGY OF THE STOMACH

XLIV. THE Origin of tHE EPIGASTRIC PAINS IN CASES OF GASTRIC AND DUODENAL ULCER

A. J. CARLSON

From the Hull Physiological Laboratory, University of Chicago

Received for publication October 20, 1917

The genesis of the pains of gastric and duodenal ulcers is not yet satisfactorily demonstrated. The view that the pains are due to acid irritation of hyperexcitable nerve endings or exposed nerve fibers in the ulcer area has probably the greatest number of adherents at present. The essential facts in support of this view are: 1. Analogy. Acid irritation of abraded areas of the skin, mouth or nose epithelium, etc. produces pain. 2. The frequent occurrence of so-called gastric hyperacidity in gastric and duodenal ulcers. 3. The temporary alleviation of the ulcer pain by food and alkalies.

The following facts appear to run counter to, or are at least not readily explained by this acid corrosion theory:

1. Gastric ulcer with or without clinical hyperacidity1 may be present without pain.

2. Gastric ulcer and ulcer pain may be associated with normal acidity, and even with hypoacidity.

3. The pains of gastric ulcer may be present and be temporarily relieved by food or alkalies, even though the stomach contents are alkaline (1).

4. Introducing acids (0.5 per cent HCl) into the stomach does not, or at least not invariably, induce or augment the ulcer pains in gastric ulcer patients (7).

1 While there is no evidence that Alexis St. Martin at any time had acute or chronic ulcer of the stomach or duodenum, Beaumont on several occasions observed raw patches on the gastric mucosa from which blood or pus exuded. In most instances this condition of the mucosa did not cause pain or discomfort, even though the abrasions of the mucosa persisted for several days. Beaumont remarks (p. 240): “It is interesting to observe to what extent the stomach may become diseased without manifesting any external symptoms of such disease."

5. The ulcer pains usually show a periodicity (being described as "gnawing" or "boring"), and the periods are too short to be explained by variations in the gastric acidity.

The other leading view ascribes the ulcer pains to contractions of the stomach, the pylorus and possibly the upper end of the duodenum, the excessively painful character of these contractions in ulcer cases being due either to hyperexcitability of the gastric pain nerves or to abnormally strong contractions. This theory has been fortified by clinical observation and experimental data especially by Hertz (7), and striking confirmatory findings on ulcer patients have recently been reported by Ginsburg, Tumpowsky and Hamburger (5). The following facts appear to support this theory:

1. Strong contractions or a certain type of contractions of the stomach and intestines give rise to varying degrees of pain in the absence of ulcer ("hunger pangs," "colic," tenesmus, various forms of gastralgia, etc.).

2. Pain nerves appear to be absent from the gastric mucosa (4). 3. The evident synchrony of the ulcer pains with gastric and possibly pyloric contractions, so far as this phase has been studied by the balloon and X-ray methods (4), (5), (7), (13), (14).

4. The frequent association of typical gastric ulcer pains with appendicitis, cholecystitis and even achylia gastrica, (6) not complicated with gastric ulcer.

5. The similarity of the moderate ulcer pains as regards periodic exacerbation with the strong hunger pangs of normal persons, a similarity that lead Moynihan and others to designate the pains of gastric and duodenal ulcers as "hunger pains."

But while the gradually accumulating reliable data thus point to gastrointestinal contractions as the immediate cause of the ulcer pains, we do not know: (1) what part of the digestive tract (stomach, pylorus or duodenum) is primarily concerned; (2) whether the site of the painful contraction varies with the location of the ulcer; (3) or what is the relation of gastric acidity to the initiation of the contractions. The observations here reported were undertaken in the hope of securing clearer knowledge of these factors.

Mr. G. H. M., age 25, the subject in these studies, was at the time a medical student in our laboratory. Six months before coming under our observation he suffered an acute attack with all the typical symptoms of "peptic ulcer," including hematemesis. After several weeks in a hospital on medical treatment and an ulcer diet, he recovered suf