As Mering and Minkowski[126] first showed, every dog whose pancreas is extirpated invariably becomes diabetic, and this diabetes is similar to that of man. In many cases of diabetes changes in the pancreas have been found at autopsy; and although a good number of cases without any apparent change in the pancreas have been recorded, the cause of these has been revealed by an American author, Dr. Opie,[127] then of the Johns Hopkins University in Baltimore. This author has found in a few cases of diabetes alterations in the islands of Langerhans in the pancreas. This fact has been confirmed by a good many authors, of whom I may mention Sobolew,[128] Weichselbaum[129] and Stengel, Sauerbeck, and others. As is invariably the case, the statements of these authorities have been attacked by others, as Hanseman, who have found no such changes in the islets in diabetes. We must, however, state here, that an apparently perfect anatomical condition of glandular structure after death need be no proof of a perfect secretory activity during life. Every epithelial formation, and the islands of Langerhans are of this nature, must furnish a secretion, and this flows in every gland only under a nervous stimulation. As Pawlow has shown, the pancreas secretes under nervous impulse. Therefore the findings of pathological anatomy cannot show us whether these glands have been secreting properly during life or not, especially in a nervous disease such as diabetes. Footnote 126: Mering und Minkowski: Archiv für exper. Path. und Pharm., xxvi, 1889. Footnote 127: Opie: Journal of Experiment. Medicine, p. 827, 1901. Footnote 128: Virchow’s Archiv, clxviii, p. 1. Footnote 129: Wiener klin. Wochenschrift, 1901-1902. According to the experiments of Diamare and Kuliabko, the islands of Langerhans furnish a secretion which aids in the inversion of grape sugar. We have shown at meetings of the Biological Department of the Hamburg Medical Society (Germany), and of the London Pathological Society, specimens of the pancreas of diabetic persons where there were changes only in the islands of Langerhans and none in other parts of the pancreas while, in one case, nearly the whole of the pancreas was destroyed by cirrhosis, but without any diabetes, for in this case the islands of Langerhans were not changed at all. It is interesting to note that these changes in the islands of Langerhans were also present in mild cases of diabetes. Thus, the finding of sugar in the urine usually indicates the probability of changes in the pancreas. Even small quantities of sugar, such as appear only after starchy meals (alimentary glycosuria _ex amylo_) may be brought into connection with changes in the pancreas; for Wille[130] found in Hamburg from a large series of autopsies, that in a considerable number of cases with alimentary glycosuria, there were also changes in the pancreas. Footnote 130: Quoted after Professor Oser: Deutsche klinik., vol. v, p. 158.