ADAPTATION OF PANCREAS. 95 



pancreatic infusions was not excluded, and therefore one can not rely 

 implicitly on the fermentation test applied by him. 



Later experiments' 61 on feeding with galactose instead of lactose also 

 produced a small increase in rotatory power, but only a slight evolution 

 of gas on fermentation with Sacch. apiculatus, but in this case Weinland 

 considered that his result was negative. 



In Bainbridge's experiments most of these sources of error were 

 avoided. In almost all his experiments a control experiment was made 

 with boiled juice or boiled infusion. He rightly discarded the polari- 

 metric method, and adopted as his criterion the reducing powers of the 

 solutions as determined by Pavy's solution. The osazone method 

 employed by him was unsatisfactory, since in no case was there isolation 

 of the osazone or determination of its melting point and its nitrogen 

 content. Moreover, although Pavy's method can be applied with 

 considerable accuracy to the determination of reducing sugar, its accuracy 

 depends on the scrupulous observance of identical conditions in all the 

 determinations. Small alterations in the time of boiling or in the 

 volume of fluids used may produce differences in the results, which are 

 serious when, as in Bainbridge's experiments, deductions are to be 

 drawn from differences in the reducing power of two solutions amounting 

 to only *2 to '9 c.c. in a total quantity of fluid of between 4 and 8 c.c. 



The method employed by Bierry consisted in the preparation of the 

 osazones after separating proteid-subs-tances by means of mercuric 

 nitrate. He considered that this was the most delicate method, but 

 found that it was inaccurate unless 20 per cent, of lactose were hydrolysed. 

 Unfortunately no data are given of the osazone obtained, except that it 

 was soluble in a mixture of equal parts of acetone and water, in which 

 glucosazone is insoluble. The possibility that less than 20 per cent, 

 inversion had occurred is not excluded, since the length of time during 

 which the ferment was allowed to exert its action is not stated. 



In my reinvestigation of the subject I have endeavoured to avoid 

 the above sources of fallacy. Of the various methods of determining the 

 question of inversion there is no doubt that, as Bain bridge has pointed 

 out, a comparison of the reduction before and after conversion is the 

 most delicate. 



Of the methods for determining the reducing power of a solution, 

 practically the only one which is free from sources of error, depending on 

 the experimenter, is that in which the copper oxide produced is collected 

 and weighed. I have therefore made use of Allihn's method for the 

 determination of the reducing power of the solutions. Moreover, before 

 PH. xxxiv. 7 



