187 
L 9 15-1 6.] Estimation of Sugar in the Blood. 
5 c.c. of the supernatant fluid is used in the sugar estimation. The sugar is 
estimated by a modified Bertrand process in which a known quantity of 
glucose is added to the copper solution. The cuprous oxide is filtered off 
through asbestos, washed, treated with acid ferric sulphate solution, and 
the amount of sugar calculated from a permanganate titration. 
(2) Recently MacLean has published still another method for the 
estimation of blood sugar in which 1 c.c. of blood is used. The method 
for removing the protein is that employed by Gardner and MacLean, 
the only difference being that a small quantity of phosphoric acid is 
added to hasten filtration. The total bulk of fluid before filtration is 
40 c.c., and 20 c.c. of the filtrate are used in the subsequent estimation. 
The glucose solution is boiled for 10 minutes with an alkaline copper 
solution containing potassium iodide and a known amount of potassium 
iodate, a reflux condenser being used to prevent reduction in bulk. The 
flask and its contents are cooled and a measured amount of pure con- 
centrated hydrochloric acid added. The iodine thus liberated oxidises 
the cuprous salts to cupric, and the excess of iodine is estimated by means 
of a standard thiosulphate solution. 
(3) According to the method of Ivar Bang, about 100 mg. of blood 
are soaked up into a small piece of specially prepared filter paper, 
which is weighed before and after on a torsion balance. The paper is 
transferred to a clean, dry, rather wide test tube, and 6*5 c.c. of a 
boiling solution of potassium chloride containing a trace of hydrochloric 
acid are poured upon it. By this means the albumen is coagulated and 
the sugar diffuses into the surrounding solution. After half an hour 
the liquid is poured into a small Jena flask, the flange of whose neck 
has been cut off, the paper is washed with 6*5 c.c. of the potassium chloride 
solution, and the washings added to the contents of the flask. 1 c.c. 
of a solution containing 4*4 gm. CuS0 4 , 5H 2 0 ; 160 gm. KHCO s ; 
100 gm. K 2 C0 3 ; and 66 gm. KC1 made up to a litre with boiled-out 
distilled water is added. A piece of thick- walled rubber tubing is fitted 
over the neck of the flask and the contents heated over a flame so 
adjusted that boiling begins in from 1 minute 25 seconds to 1 minute 
35 seconds. Boiling is continued for exactly 2 minutes, the rubber tubing 
is closed with an efficient clip and the flask cooled in a current of cold 
water for 1 minute. The rubber tubing is now removed and the cuprous 
N 
chloride estimated by titration with ^qq iodine from a micro-burette, 
starch solution being used as an indicator. To prevent oxidation by the 
air, a current of C0 2 is passed into the flask while this titration is being 
