THE METABOLISM OF THE CARBOHYDRATES 679 



and von Mering in Germany, and de Dominicis in Italy, by thorough 

 extirpation of the gland, succeeded in producing in dogs a marked and 

 persistent glycosuria, accompanied by many of the other symptoms of 

 diabetes. The first two authors attributed the condition to removal of 

 an internal secretion. 



The course of the diabetes thus produced is, however, somewhat differ- 

 ent from that usually observed in man. It is extremely acute from the 

 start, the G: N ratio being 1 :3.6 (see page 664), and it is unaccompanied 

 by any of the classical symptoms seen in the clinical condition. Experi- 

 mental pancreatic diabetes can, however, be made to simulate very closely 

 the disease in man. This was first of all demonstrated by Sandemeyer, 

 who found that if the greater part of the pancreas was removed, the 

 animals for some months, if at all, were only occasionally glycosuric, 

 but later became more and more frequently so, until at last the condition 

 typical of complete pancreatectomy supervened. Similar results have 

 more recently been obtained by Thiroloix and Jacob, in France, and by 

 Allen in this country. These investigators point out that different re- 

 sults are to be expected according to whether the portion of pancreas 

 which is left does, or does not, remain in connection with the duodenal 

 duct. When this duct is ligated, atrophy of any remnant of pancreas 

 that is left is bound to occur, and this is associated with rapid emacia- 

 tion of the animal, diabetes and death. When the remnant surrounds a 

 still patent duct, a condition much more closely simulating diabetes in 

 man is likely to become developed one, namely, in which there is, for 

 some months following the operation, a more or less mild diabetes, 

 which, however, usually passes later into the fatal type. 



It is, of course, difficult to state accurately what proportion of the 

 pancreas must be left in order that the above described condition may super- 

 vene. Leaving a remnant amounting to from one-fifth to one-eighth 

 of the entire gland is commonly followed by a mild diabetes, whereas 

 if only one-ninth or less is left, a rapidly fatal type develops. As in 

 clinical experience, the distinguishing feature between the mild and the 

 severe types of experimental pancreatic diabetes is the tolerance toward 

 carbohydrates. In the mild form, no glycosuria develops unless carbo- 

 hydrate food is taken; in the severe form, it is present when the diet is 

 composed entirely of flesh. It is thus shown that "by removal of a 

 suitable proportion of the pancreas, it is possible to bring an animal 

 to the verge of diabetes, yet to know that the animal will never of itself 

 become diabetic. . . . Such animals, therefore, constitute valuable 

 test objects for judging the effects of various agencies with, respect to 

 diabetes" (Allen 18 ). It therefore becomes theoretically possible to in- 

 vestigate, on the one hand, other conditions which will have an influence 



