32 



THE CAKBOHYDEATE ECONOMY OF CACTI. 



40ox 



steps if the cuprous oxide is dissolved. Furthermore, the filtration must 

 proceed as rapidly as possible and not infrequently do these contaminating 

 substances greatly delay this process. Thus there is always danger of 

 resolution of the cuprous oxide in the alkaline liquid, especially in the 

 presence of dissolved atmospheric oxygen. It was therefore desirable to 

 employ a method which did not depend upon the determination of the 

 cuprous oxide formed in the reduction and which also obviated the necessity 

 of filtering the solution. The only alternative was to employ the residual 

 method, which depends upon the estimation of the unreduced copper. The 

 accuracy of such a procedure naturally depends upon a good method of 

 determining copper. This has been worked out by Peters in the iodide 

 method and yields most satisfactory results. 



In the method used in this investigation, the reduction is 

 carried out in specially designed tubes (fig. 1) of a capacity of 

 10 to 50 c. c., depending upon the amount and concentration of 

 the sugar solution to be determined. A very liberal excess of 

 copper over that required to oxidize the sugar in solution is 

 / ^ ^ always provided. After heating a definite length of time, the 

 |r ^|| | tubes are rapidly cooled, made up to volume with distilled water, 

 shaken in order to mix thoroughly, and then centrif uged until all 

 the cuprous oxide (and other solid material) is in a compact mass 

 at the bottom of the tube. The residual copper in the clear 

 supernatant solution is then drawn off with a pipette and several 

 determinations are made by the iodide-thiosulphate method. The 

 difference between this result and the original strength of the 

 copper solution represents the amount of copper reduced and is 

 a quantitative index of the amount of sugar originally present. 

 Experience with this and various other methods of sugar deter- 

 mination by means of alkaline copper solutions has taught that 

 reliable and quantitatively comparable results can be obtained 

 only by means of a most careful quantitative standardization of 

 each step in the manipulation. Hence it is also necessary to adhere to the 

 same procedure throughout an investigation. The sugar determinations 

 were made by use of Fehling's solution (Soxhlet's modification: 69.28 

 CuSO 4 5H 2 O per liter), although it was later found that a solution con- 

 taining sodium carbonate instead of caustic alkali possessed certain advan- 

 tages, because the weaker alkali affects the sugar less profoundly. The 

 Fehling's solution was used throughout in order to maintain the uniformity 

 of procedure. Thus solution A was made up of 69.28 grams of pure 

 CuSO 4 5H 2 O in 1,000 c. c. water, while solution B contained 346 grams 

 of sodium potassium tartrate and 100 grams of sodium hydroxide in 1,0< 

 c. c. of water. As will be seen later, a correction for the spontaneous reduc 

 tion which takes place in this mixture was automatically adjusted in the 

 blanks and standardization by means of pure d-glucose. 



FIG. 1. 



