should be noted here that accompanying the hypoglycemia in normal rabbits, the glycogen of the liver is discharged and converted into glucose in the attempt apparently to bring the blood sugar content back to normal. The reason for the hypoglycemia is then to be sought in disappearance of sugar from the muscles. How does insulin act to decrease the sugar of the tissues? It has been suggested that insulin acts to increase the rate of oxidation of glucose in the tissues. Experiments with normal animals in which the respiratory quotient, that is, the ratio of carbon dioxide produced to oxygen consumed in an animal, has been determined, have given contradictory results. Certainly the disappearance of glucose from the tissues can not be explained entirely by oxidation. Nor is the lost glucose converted into glycogen and stored in the muscles. A third possibility is that glucose is transformed in the presence of insulin into some substance, as yet unrecognized, which has lost the characteristic properties of glucose and which is presumably more easily oxidized, possibly y-glucose, hexose phosphoric acid, or lactic acid. This theory which appears most plausible, while formerly more or less discredited by failure to confirm some of the earlier experiments on which it is based, is now again becoming of interest in the light of more recent work.

The influence of insulin administered to depancreatized dogs is striking. It will be recalled that in such animals, no glycogen is stored as a reserve carbohydrate, glucose is not burned, hyperglycemia occurs and the excess sugar appears in the urine. In

comparing the effect of insulin on normal animals with that obtained in depancreatized dogs, it is necessary to remember that in the former, we are superimposing the effect of extra insulin on an organism already adequately supplied with sufficient amounts, while in the latter we are merely making good an existing deficiency. After the injection of insulin into such depancreatized animals, the excretion of glucose in the urine ceases, the hyperglycemia rapidly disappears. If carbohydrate be fed, glycogen is stored, combustion of sugar occurs, and the untoward effect of failure to combust glucose as evidenced by failure to properly burn fats, resulting in acidosis no longer is evident. The important point is that in animals with an insulin deficiency occasioned by removal of the pancreas, administration of insulin results in a return to nearly normal conditions as far as concerns the combustion of fat and carbohydrate. A similar condition exists when insulin is administered to diabetics in whom pancreatic insufficiency has produced disorders of carbohydrate and fat metabolism. The results with diabetic patients are, however, more difficult to predict with certainty, due to the fact that in these, complete failure of pancreatic function is not present and a therapeutic dose of insulin is superimposed upon a certain amount, although insufficient, present already.

The effects of insulin are not permanent and disappear rapidly. As to its fate, we cannot speak with certainty. Some is apparently excreted in the urine, some exists in all tissues, although in much smaller amounts than in the pancreas.

The occurrence of insulin in animals other than mammals has served to furnish an interesting confirmation of the theory that the site of formation of insulin is in the islets of Langerhans of the pancreas. It will be recalled that mammalian pancreas contains two types of cells, the acinous or zymogenetic, furnishing the enzymes of pancreatic juice, and the islets of Langerhans. In certain fish, as for example, the sculpin and flounder, the islets are relatively large and so separated from the zymogenetic tissue as to make a macroscopic separation of the two types of cells possible. This fact, discovered much earlier by Diamare, made possible proof of the theory of localization of insulin formation. Extracts of islets and of the acinous tissue of the flounder were made by the method outlined and tested. The results were clean cut and showed conclusively that the islets give large yields of insulin and that this is absent from acinar tissue. The practical importance of this fact has been emphasized by the workers of the Toronto laboratories. In the cod and halibut, the principal islet is large and so situated as to be easily recognizable by the unskilled fisherman. Highly potent insulin was prepared from this source and proved suitable for clinical use on diabetic patients. The yield of insulin obtained, expressed per kilogram of tissue in order to facilitate comparison with the yields from beef pancreas, was from 14 to 22 thousand units. When one considers that the islets were removed by unskilled labor from the fish after they had been brought to shore and that several days had elapsed before the preparation of the extracts, these yields are

striking and the results offer a promising source of insulin for the future.

It may be assumed that insulin is present in all animals, although in invertebrates in which structures similar to the islets have not been described, the site of production is not known. Collip, one of those originally associated with the Toronto laboratories, has prepared an insulin like substance from clams and also from yeast. He has suggested that insulin is present wherever glycogen, the storage form of glucose, is present. Substances which exert a hypoglycemic action have also been described as occurring in many plants, but the onset of the hypoglycemic effect is delayed frequently for twentyfour hours and the diminution of blood sugar is usually of longer duration. Whether they act directly on sugar metabolism or indirectly through stimulation of the pancreas is not known, nor has their therapeutic use been attempted.

We may sum up the principal facts concerning insulin as follows. An active principle is present in large amounts in the islets of pancreatic tissue and can be successfully extracted, provided the destructive action of the products of the digestive cells of the pancreas is prevented, usually by the action of fairly high concentration of acid. This active principle, insulin in purified form, exhibits the following characteristic properties. When given to normal rabbits, it causes immediate reduction of the blood sugar and when hypoglycemic symptoms supervene these can be immediately and permanently removed by glucose injections. When given to animals rendered diabetic

by removal of the pancreas, the symptoms, hyperglycemia, acidosis, and glycosuria should immediately disappear; the respiratory quotient rises and glycogen is stored in the liver if carbohydrate is also fed.

It is the purpose of this discussion to present insulin from the chemical and physiological point of view. A brief discussion of the results of the therapeutic administration of insulin to diabetics as summarized by Banting, himself, may, however, be added.

Insulin enables the severe diabetic to burn carbohydrate as shown by the rise in the respiratory quotient following the administration of glucose and insulin. It permits glucose to be stored as glycogen in the liver for future use. The burning of carbohydrate enables the complete oxidation of fats, and acidosis disappears. The normality of blood sugar relieves the depressing thirst and consequently there is a diminished intake and output of fluid. Since the tissue cells are properly nourished by the increased diet, there is no longer the constant calling for food, hence hunger pain of the severe diabetic is replaced by normal appetite. On the increased caloric intake, the patients gain rapidly in strength and weight. With the relief of the symptoms of his disease, and with the increased strength and vigor resulting from the increased diet, the pessimistic, melancholy diabetic becomes optimistic and cheerful.

Insulin is not a cure for diabetes; it is a treatment. It enables the diabetic to burn sufficient carbohydrates, so that proteins and fats may be added to the diet in insufficient quantities to provide energy for the economic burdens of life.-International Clinics, December, 1924.

In conclusion, it should be pointed out, without in any way detracting from the brilliant work which resulted in the discovery of insulin, that proper utilization of this discovery was made possible only as a result of the work of countless other investigators on carbohydrate metabolism and diabetes. The stage was set, so to speak, for insulin. Without the knowledge of intermediary carbohydrate metabolism and dietary control of diabetes which antedated the discovery of insulin, the clinical value of insulin would have been slight. It has been said with considerable truth that had the discovery been made ten years earlier, its clinical application would have been more difficult because we then possessed no simple method of blood sugar estimation on which the standardization and control of the clinical use of insulin depends to so large an extent. Epinephrin, the active principle of the adrenals, was isolated in 1901, a discovery equalling in its brilliancy the discovery of insulin. We knew at that time and still know little with certainty of the exact function of the adrenals and adrenaline has little value therapeutically except as a drug acting upon the vaso-motor system. Insulin, discovered in 1922, the active principle of a gland concerning whose effects on metabolism we already knew much, has proved of the utmost value therapeutically.

H

ALDANE (1923) has recently reported that large doses of sodium bicarbonate (0.6 gram per kilogram) when fed to normal subjects receiving normal diets caused an excretion of the acetone bodies; simultaneously there occurred a decrease in the combustion of carbohydrate shown by a marked lowering of the respiratory quotient. This form of ketosis is therefore comparable with that which has been repeatedly described when carbohydrate is withdrawn from the diet or when the body is, for some reason, unable to burn that food. It has been thought worth while to investigate the effect of moderate doses of sodium bicarbonate such as are used in mild alkali therapy upon the excretion of the acetone bodies.

Effects of sodium bicarbonate upon an established ketosis have often been studied, and the results reported by different observers have frequently been more or less conflicting; only a

A preliminary report of part of this work was read before the Western New York Branch of the Society for Experimental Biology and Medicine in Rochester, New York, October 11, 1924.

few of these will be cited as illustrations. Joslin (1904) reported the results of a series of short fasts upon a normal person. He found that the rate of excretion of the acetone bodies was approximately the same during two days of starvation if sodium bicarbonate was or was not ingested. In the discussion of his results he pointed out that as a rule alkali had been found to increase the rate of excretion of the acetone bodies. Forssner (1911) found an increased excretion of acetone when he added sodium bicarbonate to a diet low in carbohydrate which had been used to induce a condition of ketosis in a normal subject, and Hubbard and Wright (1922) found similar results under similar metabolic conditions. ditions. Joslin in his text book has pointed out the frequency with which such an increase is met with in the treatment of diabetes, and Allen, Stillman, and Fitz3 have shown that such results are not found in all cases. In general, however, there is no doubt that an increased excretion of acetone does follow the administration of

* Joslin (1917), pp. 394 and 395. Allen, Stillman, and Fitz (1919), pp. 109-115.

alkali. Mosenthal and his coworkers (1922, 1923) in experiments not yet published in full, have stated their belief that sodium bicarbonate fed to a diabetic patient causes a washing out of the acetone bodies in the blood rather than an increased production of these compounds. Such an explanation cannot apply to the experiments of Haldane, for there is practically no acetone in the blood of normal subjects receiving a normal diet, but it may explain the effect of moderate amounts of alkali upon the acetone excretion of subjects in whom a ketosis already exists.

It was thought that if subjects were brought just to the border-line of ketosis and alkali administered to them while they were in this condition, the effect of smaller amounts of sodium bicarbonate than previous observers have studied might be shown, and that the manner in which the alkali affects the excretion might be decided. Such experiments are attended with considerable difficulty. Slight variations in diet-particularly in the antiketogenic foods-and marked changes in the metabolism such as can be brought about by significant differences in the amount of exercise may affect the results. A conservative interpretation should therefore be placed upon such studies.

Brief protocols describing all of the cases-5 in number-and tables containing the results of determinations of the acetone bodies in urine, done by a method previously described (Hubbard, 1921), are presented below. All were cases of arthritis, as were those upon which previous results have been obtained (Hubbard and Wright, 1922). The diets were calculated to meet the

needs of each of the patients, and in most instances contained more calories than were necessary for their metabolic requirements. Two of the patients gained weight during the experiments. The diets used were taken from the tables compiled by Clarke (1924) for use in the treatment of arthritis, and were based upon the tables of the composition of foods of Bryant and Atwater (1906). They were fed in three meals and as far as was possible a third of the protein, fat, and carbohydrate was given in each meal. The sodium bicarbonate, when it was used, was given in three equal doses, one after each meal.

Since no attempt has been made to interpret the diets in terms of their ketogenic balance, differences between the calories ingested and burned are relatively unimportant. The question of the constancy of the ketogenic: antiketogenic ratio in any given diet however is important, and the probability of its maintenance must be briefly discussed. Not all of the subjects adhered equally closely to the diets and in some the differences between food eaten and food furnished were great. It was found however that it was the foods which contained the larger part of the antiketogenic material which were eaten, and the fat which was left. It has been shown that in experiments of this kind ingested fat and body fat give rise to equal amounts of the acetone bodies (Hubbard, 1923); and therefore the discrepancies in the diets were not as serious in their probable effect upon the experiments as a study of the reports from the diet kitchen at first seemed to imply. The effect of varying amounts of exercise taken by the patients in