Although, many theories as to the cause of diabetes have been offered, the concensus of opinion is that it arises either from irritation of the diabetic centre in the floor of the fourth ventricle, or from disease of the Islands of Langerhans in the pancreas. In the former case the glycogenic function of the liver is disturbed, and in the latter it is probable that there is failure of a specific internal secretion, whose particular property is to oxidize the sugar molecule, from disease of the groups of cells in the pancreas which produce it.

That the glycosuria of pancreatic diabetes is due to the failure of a specific internal secretion is suggested by Tuckett (24) who found that if thoracic lymph from a digesting dog were injected into the portal circulation of a cat, a glycemia of 1 per cent, and a glycosuria of 9 per cent, resulted, but the lymph from a fasting dog occasioned no sugar.

It has furthermore been suggested that diabetes may originate from toxic substances absorbed from the intestinal tract, and Topfer and others induced diabetes in dogs by feeding them fecal material from diabetic patients. These theories are as yet sub judice and we await more complete experimental evidence.

It has been found, however, that the coma of diabetes is due to a saturation of the alkalies of the blood by the products of incomplete combustion, and these are of an acid nature like oxy-butyric acid. Walter, (25) by the administration of acids to animals, has produced dyspnea, loss of motion, collapse and unconsciousness, - a state closely resembling diabetic coma. Upon examining the blood of animals thus poisoned he found that the amount of carbonic acid was only 2 per cent, to 3 per cent, by volume, which is the amount always dissolved in the blood, normal venous blood containing nearly 50 per cent. As the alkalies of the blood were saturated by acid, no carrier of carbonic acid remained, and this accumulated in the brain and caused its usual toxic symptoms. Minkowski (26) has shown that the amount of carbonic acid in the blood of a comatose diabetic patient was only 3.3 per cent, by volume, and the blood obtained post-mortem had a distinctly acid reaction and contained large quantities of oxy-butyric acid and sarcolactic acids.

Acetone resembles alcohol and ether in its toxic properties and can be given to dogs in dose of 1 g. per K. without any effect; 4 g. per K. cause symptoms of intoxication with motor disturbances like those of ethyl alcohol; 8 g. per K. is fatal, - hence to a person weighing 75 K., 600 g. of acetone would be necessary to cause death, an amount which could not be caused by the decomposition of proteid. We may then infer that acetone is not a prime cause of diabetic coma. Although ammonium is increased to five or ten times its usual amount in diabetes, it is probably neutralized by acids so that it does not participate in causing coma or convulsions.

The relations of B-oxy-butyric acid in diabetes have been studied clinically by Herter (27) who formulates the following important conclusions: -

1. Crotonic acid can usually be obtained from the urine of persons in diabetic coma.

2. Diabetic coma is preceded by a period of days, weeks or months in which there is a large excretion of B-oxy-butyric acid.

3. The persistence of more than 25 g. per diem indicates impending coma.

4. A patient passing 30 g. of B-oxy-butyric acid may be able to be about all day and do considerable muscular work.

5. The acid intoxication is an important element in prognosis, and frequent chemical examinations are necessary to show the results of treatment.