completely disappear in this case. After several months on a moderately careful diet she still has - per cent of sugar more or less constantly.

In considering the relation of the thyroid to glycosuria we find certain well defined facts. On the experimental side we know that the feeding of thyroid tablets to normal healthy individuals may lead to the development of a lowered carbohydrate tolerance, or to an actual glycosuria. This happens still more frequently if we give thyroid extract to goitre patients. Generally, such a glycosuria ceases with the cessation of the thyroid administration, but in rare instances it has persisted and even gone on to the development of a severe diabetes with acidosis.

Now with regard to the typical cases of exopthalmic goitre, we find glycosuria to be of fairly common occurrence. Kocher reports it in 2 per cent, Sattler in 3 per cent of his cases. This is too frequent to be a mere coincidence, and we know that some casual relationship must exist between the hyperthyroidism and the glycosuria. Clinically the glycosuria may vary from the mildest to the severest type; it may precede, be coincident with, or follow the development of the exophthalmic goitre symptoms; and its intensity is in no way related to the intensity of the Basedow. In Sattler's statistics covering 40 cases, there were 26 cases in which the Basedow developed first, six cases in which the diabetes developed first, and eight cases in which the two conditions developed simultaneously. As was stated, the glycosuria may be of any grade; so, too, the exophthalmic goitre symptoms may be of any grade. Falta has called attention to a group of cases in which an incomplete or abortive type of Basedow is complicated by glycosuria. In these cases the Basedow symptoms may be so slight as to escape recognition. Such cases would then be classed as true diabetes, and treated as such, whereas, to produce a cure, the treatment should be directed against the hyperfunctionating thyroid. And finally, there may be all degrees of dependence of the diabetes upon the hyperthyroidism. There are undoubtedly many mild cases of glycosuria which cease with the cessation of the Basedow symptoms and there are some severe cases of diabetes, as Dr. O'Day pointed out, in which the cure of the hyperthyroidism resulted in the cure of the diabetes. But just as surely there are many cases in which the operation influences the diabetes only to a certain degree, and still others in which the diabetes runs its course, even to a fatal termination irrespective of what happens to the goitre. Taking this in connection with the fact that the feeding of thyroid tablets produces glycosuria only in certain

cases, by no means in all, we are forced to the realization that the relationships which we are studying are very complicated.

Without going into a discussion of the various hypotheses which attempt to explain these phenomena, we wish merely to lay stress on the eminently practical fact, first, that in certain cases where diabetes and Basedow occur together, both may be cured by an operation for the relief of the hyperthyroidism; and, second, that far from the diabetes being a contraindication for the operation, it is actually an additional indication to operate. You will not cure the diabetes in every case; but you will cure it in some and help it in many. And that is offering a great deal when you consider the gloomy outlook given in the older statistics in these cases where hyperthyroidism and diabetes coexist.

MYXEDEMA

In Basedow's disease we have a hyperfunction of the thyroid; in myxedema on the other hand, a hypofunction of the gland. After what has been said of the relationship of hyperthyroidism to glycosuria, one would naturally expect to find in myxedema an increased tolerance to sugar. And such is actually the case, where the thyroid function has been reduced or lost, either as a result of disease or as a result of removal of the gland, we find that very large doses of sugar, 300 to 500 grams or more, may be given by mouth without the appearance of sugar in the urine. In certain of these cases thyroid medication for the myxedema has resulted in glycosuria.

But the ductless glands are full of contradictions. There are undoubted cases of myxedema, and not a few of them, which show spontaneous glycosuria. Strangely enough, this spontaneous glycosuria tends to clear up, along with the other myxedemic symptoms, under thyroid medication.

The confusion seems inexplicable, unless we can bring ourselves to realize that the diseases of the thyroid are only in isolated instances purely thyroid disease, and that in the majority of cases they represent simply a dominant thyroid element in a polyglandular symptom complex.

HYPOPHYSIS

Glycosuria occurs in about 40 per cent of all cases of acromegaly. It may vary fram an alimentary glycosuria, to an enormous daily sugar excretion. Indeed, there is no other disease in which such a large output of sugar occurs, day in and day out, extending over years. Some of these cases show the typical symptoms and complications of a diabetes mellitus. On the other hand, in many of the cases, the glycosuria per

sists as a pure symptomless complication of the acromegaly. In these cases the glycosuria may show marked variations, and in comparison with diabetes mellitus is relatively independent of the variations in diet. Cushing points out that it is chiefly in the early stages of the disease that the lowered sugar tolerance is evident. Later, as the activity of the gland becomes reduced, and symptoms of diminished hypophysial secretion appear, the tolerance rises, and may even become markedly increased.

According to Von Noorden and his co-workers, the lowering of the sugar tolerance is dependent, as is the acromegaly itself, upon an increased activity of the anterior lobe. This seems reasonable enough, when we consider the analogy to hyperthyroidism, and consider further that developmentally the anterior lobe is intimately associated with the thyroid. But there is no unanimity of opinion on this point. Cushing considered the posterior lobe responsible; others believe these glycosurias to be thyreogenic in origin; and still others believe them due to premature degeneration of the pancreatic islands. And as a final possibility which perhaps has more in its favor than any of the others, we must consider hyperactivity of the pars intermedia. Cushing's experiments could well be interpreted in this sense, and here, too, we have not only the developmental relation to the thyroid, but even a histologic similarity.

In dystrophia adiposo genitalis, or Frohlich's syndrome, we have a symptom complex depending on deficiency of pituitary secretion. Frolich's syndrome stands in relation to acromegaly much as myxedema does to Basedow's disease. In acromegaly, in the early stages at least, the sugar tolerance is lowered. As deficiency symptoms develop, as they frequently do later in the disease, the tolerance rises; and in Frohlich's syndrome, where the deficiency symptoms dominate the picture from the first, a very high sugar tolerance is the rule.

ADRENAL

Next to the pancreas, the adrenal undoubtedly plays the most important role in the control of sugar metabolism. The first step in our knowledge of this matter came many years ago, when Claude Bernard demonstrated that puncture of the floor of the 4th ventricle was followed by the appearance of a transient glycosuria. He showed further that the glycosuria did not occur if the splanchnics had been cut. The mechanism of the piqure glycosuria remained a mystery, until the discovery of adrenalin by Herter, Takamine and others, and the demonstration that the injection of very small amounts of adrenalin into the circulation resulted in glycosuria.

We now

know that when the floor of the fourth ventricle is irritated, a stimulus is carried along the sympathetics to the adrenal. This results in a discharge of an excess of adrenalin into the circulation. The adrenalin is carried to the liver, where it causes a rapid mobilization of glycogen. We then get a flooding of the blood with sugar-a hyperglycemia, which in turn is followed by a glycosuria. It is probable that a good many of the transient glycosurias, such as those produced by apoplexy, encephalomalacia, etc., act in this way through stimulation of the adrenals.

If we inject adrenalin we can produce a hyperglycaemia and glycosuria. If we If we remove the adrenals, we produce a hypoglycaemia, or fall in the blood sugar. In Addison's disease we have a hypofunction of the adrenals, as a result of a bilateral destruction of the gland due usually to a tuberculous process. And here, in addition to the classical symptoms of asthenia, gastrointestinal disturbances and abnormal pigmentations, it has been shown by Porges and others that there is a definite reduction of the blood sugar, and this has proven sufficiently constant to become one of the important differential points in the diagnosis of the disease.

While all these glands play an important role in the control of sugar metabolism and can determine transient or even persistent glycosurias, and a temporary increase or decrease in the blood sugar, still the controlling factor in the entire process is the pancreas.

This was first definitely shown by Mehring and Minkowski in 1890 when they proved that complete removal of the pancreas always produced a fatal diabetes. Later it was shown that this function of the pancreas was dependent on an internal secretion produced by the islands of Langerhans. And we now consider that, in general, the development of diabetes mellitus is dependent directly on disease of these islands. As yet it has not been passible to have a perfectly clear pathologicalanatomical proof of this point. Increasing care in microscopic examinations has brought out the fact that disturbance in size, number and apearance of the islands occur in a large percentage of cases. In general, these are of two types. The one in which the islands show hydropic degeneration-chiefly the cases of diabtees in young individuals. The other, in which the changes in the islands are secondary to sclerotic changes in the pancreas, a chronic interstitial pancreatitis.

But with the best will in the world to make clinic and pathological anatomy agree, we will meet with a not inconsiderable number of cases in which changes in the islands cannot be satisfactorily demonstrated. It is possible that in some, if

not all, of these cases, there is a functional disturbance of the islands, in spite of the absence of any gross anatomical lesions.

And now the question arises whether all cases of diabetes are explainable on this basis. In the course of our paper we have referred to the influence of the other glands of internal secretion of the thyroid, where overactivity of the gland is associated with all grades of glycosuria from the alimentary type to the severest diabetes mellitus; of the hypophyses, where overactivity of the anterior lobe may also produce a glycosuria ranging from the mildest to the severest type; and of the adrenal, where increased adrenalin production may give rise to a more or less prolonged outpouring of sugar. The cases referred to at the onset of the paper would seem ot demonstrate that these glands, and especially the thyroid, may actually initiate and determine a diabetes mellitus. But they do not actually prove the point. Even in those cases where thyroid operation resulted in the cure of the diabetes, the most that we can say positively is that in that given case the thyroid was playing the dominant role. The effect of hyperthyroidism on glycosuria is such a notoriously uncertain quantity, varying so fram case to case that we are not justified in assuming it is the sole factor in any case. After all, it may be true, as Naunyn suggested years ago, that overactivity of these glands results in diabetes, only where there is an inherent tendency to the disease—that is, where a congenital weakness of the pancreas exists. I believe that Falta is wrong when he states that it is not difficult to distinguish between pancreatogenic or thyreogenic glycosurias-for I believe that there are all transitions between the two. And, finally, whatever the truth may prove to be, we must at least realize that in interpreting diseases produced by disturbance of the ductless glands, we must take into consideration the wide variation of symptoms resulting from their interrelation and interaction.