(2) It is not desirable for all diabetics to hold their weight. Some cases may do much better if their weight is reduced ten, fifteen, or even twenty pounds. (3) After starvation, the diet must be raised very slowly, to prevent recurrence of glycosuria. (4) An excess of protein must be regarded as producing glycosuria and an excess of fat ketonuria, and the protein and fat intake must be restricted a good deal more than has usually been the custom in treating diabetes. Case Reports. It is thought worth while, for the sake of illustration, to include a few case reports. The adults were treated at the Massachusetts General Hospital, the children at the Children's Hospital. Two charts are kept for each case: one a food chart, with the amounts of the different articles of food taken each day, and the protein, carbohydrate, fat and caloric value figured out for each foodstuff; the second (see below) a more general chart, which shows graphically the progress of the case. The first three are cases which were treated first with the old method of _gradually_ reducing the carbohydrate intake and could never be made sugar-free, running from 0.1% to 0.2% of sugar. On the new treatment they responded promptly and were discharged sugar-free. * * * * * Case 1. A woman of 64, diabetic for two years. She was sent in from the out-patient department, where she had been receiving a diet of 50 grams of carbohydrate and 50 grams of protein. On this diet she was putting out 8 grams of sugar a day with moderately strong acetone and diacetic acid reactions in her urine. When the carbohydrate was cut in the ward to 30 grams, she put out 3 grams of sugar a day. She complained of severe pruritus vulvae. After sixteen days of this treatment she continued to put out from 0.1% to 0.2% of sugar a day. Allen's treatment was then started, and after one day of starvation she was sugar-free and remained so for four days on a diet of carbohydrate, 20 grams; protein, 30 grams; fat, 150 grams. The itching had gone. Then the protein was raised to 80 grams, with the carbohydrate at 20 grams, and she immediately showed 1.5% of sugar. This is very important; the protein should not be raised too quickly. This we did not realize in our earlier cases. A second starvation day, followed by two vegetable days, and a more careful raising of the diet--as follows--kept her sugar-free, and she was discharged so. Her diets were: Dec. 12. Carbohydrate, 20 grams. Protein, 30 grams. Fat, 150 grams--1500 calories. No glycosuria. Dec. 15. Carbohydrate, 30 grams. Protein, 30 grams. Fat, 200 grams--2000 calories. No glycosuria. Dec. 20. Carbohydrate, 30 grams. Protein, 40 grams. Fat, 180 grams--2000 calories. No glycosuria. Dec. 26. Carbohydrate, 40 grams. Protein, 40 grams. Fat, 180 grams--2000 calories. No glycosuria.