METHODS OF ESTIMATING CARBOHYDRATES 47 



found to be accurate and rapid. Kendall (1912) tried to 

 adapt this solution for gravimetric determinations, but 

 found it unsatisfactory. As it seemed to him desirable 

 to obtain a reagent of a more stable character than the 

 ordinary Fehling solution as used by Brown, Morris, and 

 Millar, also by Munson and Walker, and one which would 

 give a larger yield of cuprous oxide for the weight of glu- 

 cose employed, series of trials were carried out. A 

 solution of copper sulphate, potassium carbonate, and 

 salicylic acid, was found to be very suitable, as it showed 

 no reduction even after seven hours at 100. The amount 

 of copper reduced by Kendall's method is about half as 

 great again as with Fehling's solution. Furthermore, the 

 solvent action of the carbonate upon cuprous oxide is very 

 much less than that of the hydroxide. Kendall prepared 

 tables for glucose, invert sugar, lactose, and maltose. 

 These have been supplemented by one for fructose con- 

 structed by Atkins and Wilson. 



Kendall's solution is, I believe, capable of giving very 

 accurate results. For example, a certain weight of pure 

 glucose, of known moisture content, afforded a figure for 

 the weight of cupric oxide, which differed by 0-0002 gramme 

 from that recorded in the table. This was obtained with- 

 out any previous experience of the technique, at a first 

 trial. It is not advisable, however, to employ this solution 

 in cases where citric acid has been used to invert sucrose, 

 for low results are always obtained. These, however, are 

 strictly comparable inter se. 



THE LIMITED OXIDATION OF SUGARS BY BROMINE. 



It has recently been shown by Nef (1914) that ketonic 

 and aldehydic sugars may be separated by the action of 

 bromine in the cold. In this way he states that the latter 



