The mildest cases are discoverable by the disappearance of sugar from the urine before all carbohydrates are withdrawn, and in such cases the strict exclusion of this class of foods will not be necessary. In most cases of diabetes a certain amount of carbohydrate food may be eaten and will be utilized by the tissues. What this amount is can be determined only by experiment. Moreover, the ability of the same individual to utilize such food varies from time to time. It is rare that more than 100 grams (31/3 ounces) of carbohydrate in a day can be taken with safety.

Those cases that belong to more severe types of the disease cannot be so quickly classified. The effect of more rigorous dieting and often also of medicinal treatment must be observed. That diet which, in the mildest cases, has been found to prevent glycosuria, must be continued as long as possible, or at least for several weeks. When cases are very mild, the dietetic limitation necessary is often so slight as to cause the patient no discomfort and to require little self-denial. But in those cases that approach the group of moderate ones not only does the persistent restriction necessary demand an irksome self-denial, but the excess of albuminoids and fats often lessens appetite greatly and causes indigestion.

It should not be thought that a patient is cured because glycosuria does not return while he is upon a restricted diet. A genuine cure is rarely effected. Temporary recovery often occurs. When a restricted diet has been used for from six to twelve weeks, it may gradually be made more generous as regards carbohydrates. The mildest cases may be permitted a thin slice of bread at each meal, and the amount of this food may be increased for two or three days, and its effect ascertained by urinalysis. If it does not cause glycosuria, a baked potato may also be tried, and later a small quantity of some fruit that contains the least amount of sugar. It is only in a very small number of cases that a full normal diet can be resumed after several months of treatment. Recurrences are so usual that it is necessary to enjoin perseverance in watching the urine. If it increases in amount, and especially if it produces a feeling of stickiness when it dries upon any object, the patient should at once consult his physician. If appetite grows less, or weight of body diminishes, he should also place himself under guidance. It is best for all who have had glycosuria to have a urinalysis made from every four to six weeks for a year or two even after apparent recovery. In most cases recurrences take place from time to time, requiring careful treatment.

If the diet that is necessary in moderate cases to prevent glycosuria cannot be long maintained, it is still less feasible in the severest cases, in which all sugar and carbohydrate food must be excluded, necessitating a diet exclusively of meat and fat. Such a diet soon causes a disgust for the foods permitted the patient, or dyspepsia, and often increased nitrogenous denutrition, which is already excessive. By a reasonable restriction of diet the percentage of sugar in the urine can be kept low. It must be remembered that sugar in the urine per se is not harmful. It is of importance only as it indicates the degree of disturbance of nutrition that exists. When, by suitable treatment of a moderately severe or a severe case, not more than 500 grains of sugar are voided daily, the case may be regarded as well controlled.

Excellent results can often be had by intermittently restricting the diet closely. From two to six times a year this should be done for from two to four weeks. In all cases it is well, on one or two days in every week, to place patients upon a diet limited in quantity and much restricted in character. For instance, if a patient is permitted a liberal amount of meats, nonamylac-eous vegetables, and a little bread, or occasionally a potato, it is well once in every five to seven days to prescribe a day of fasting, when only a modicum of albuminous food, a little salad, and other green vegetables are permitted. Especially at these times water should be drunk freely and if possible to some extent an alkaline water. These fast days and the longer periods of restriction previously referred to remove for a time, or greatly lessen, the excretion of sugar in the urine. In prescribing these days and periods of fasting one must remember what has already been said of the significance of Gerhardt's test and of disgust for food and of dyspepsia.

To prevent too rapid nitrogenous denutrition, fats must be urged upon diabetics. They lessen the consumption of tissue albumin just as carbohydrates do. Unfortunately, they are not tolerated in quantities that will enable them fully to take the place of carbohydrates. Butter is well digested and should be eaten freely. Olive oil should be used generously upon salads. Fat meats, such as ham, bacon, pork, and mutton, should be eaten. If these meats must be eaten often they are best tolerated when eaten cold, except crisp fried bacon. Many clinicians prescribe cod-liver oil also, but this is not readily taken by patients. It is likely to cause indigestion and disgust and can rarely be taken in sufficient quantities to make it very useful. Many think that a small quantity of some alcoholic beverage that does not contain sugar will stimulate the stomach and facilitate the removal of fats from it, and therefore assist their more rapid digestion. I know of no experimental demonstration of this. It is undoubtedly true that many patients can take fats and oils in considerable quantities with more relish if they take also such an alcoholic beverage. To some extent condiments help in the same way.

Meats, except liver, oysters, mussels,1 and such shell-fish as consist chiefly of liver, may be permitted at all times, but they should not be cooked with flour or bread-crumbs. Eggs are especially useful. They contain 13 per cent, of protein and 11 per cent, of fat. Meat, eggs, and fish must constitute the staple of a diabetic's diet. Cheese of all kinds is permissible. Some kinds contain a goodly amount of fat in an agreeable form. Butter contains 83 per cent, of fat, 0.7 per cent, of sugar, and 0.86 per cent, of protein. It is, however, the most palatable of all fats, and can be taken in very large quantities with the greatest pleasure. It should be spread thickly upon bread or such substitutes for bread as may be prescribed. Cream also is agreeable, and may be permitted with benefit to most patients, although it does contain about 4 per cent, of lactose. Its 16 per cent, of fat makes it most desirable, and diabetics can often dispose of considerable quantities of milk-sugar, just as they can of levulose. Cream is often made more palatable by-dilution with water, perferably such an alkaline water as vichy.

1 Kleen and others permit shell-fish; oysters contain only 2.6 per cent, of carbohydrate.