28 THE ROYAL SOCIETY OF CANADA 



deterioration of the potency of the extract. The alcohol could also 

 be removed by distillation in vacuo at low temperature and it was 

 found by reducing the volume to one-fifth instead of to dryness 

 that a watery extract of the active principal was obtained. This 

 could be sterilized by passing it through a Berkfeld filter. The 

 extract in this form was given to a human diabetic and results in 

 every way comparable to those obtained on the depancreatized dog 

 were observed. However, owing to the high percentage of protein 

 also present, sterile abcesses formed in a few instances at the site of 

 injection. 



2, The Preparation of the Extracts as used in the first Clinical Cases 



(J. B. COLLIP) 



The demonstration by Banting and Best that extracts of pan- 

 creas, prepared with certain precautions, contain a substance having 

 the power to lower the blood sugar and to raise the sugar tolerance 

 in diabetes, both in dogs and man, warranted an attempt to isolate 

 this substance in a sufficiently pure state for repeated subcutaneous 

 or intravenous administration in man. The problem was to remove 

 most of the protein and salts and all of the lipoid material from the 

 extracts without destroying the active principle. In the endeavour 

 to solve this problem various methods were tried and the following 

 was found to be most satisfactory. 



To a small volume of 95 per cent, ethyl alcohol freshly minced 

 pancreas was added in equal amount. The mixture was allowed to 

 stand for a few hours with occasional shaking. It was then strained 

 through cheese cloth and the liquid portion at once filtered. The 

 filtrate was treated with two volumes of 95 per cent, ethyl alcohol. 

 It was found by this treatment that the major part of the protein 

 was removed while the active principle remained in alcoholic solution. 

 After allowing some hours for the protein precipitation to be effected 

 the mixture was filtered and the filtrate concentrated to small bulk 

 by distillation in vacuo at a low temperature (18° to 30°C). The 

 lipoid substances were then removed by twice extracting with sul- 

 phuric ether in a separating funnel and the watery solution returned 

 to the vacuum still, where it was further concentrated till it was of 

 a pasty consistency. 80 per cent, ethyl alcohol was then added and 

 the mixture centrifuged. After centrifuging, four distinct layers 

 were manifested in the tube. The uppermost was perfectly clear and 

 consisted of alcohol holding all the active principle in solution. Below 

 this, in order, were a fiocculent layer of protein, a second clear watery 



