189 
1915-16.] Estimation of Sugar in the Blood. 
of the utmost importance. Various kinds of clips were tried, but a pair 
of surgical artery forceps was found to be much the best instrument for 
this purpose. A very rapid current of cold water must be used for cooling 
the flask and its contents, and care must be taken to insert the leading 
tube from the C0 2 generator into the flask as soon as the rubber tubing 
is removed from the neck. The starch solution may afterwards be added 
N 
by means of a small pipette pushed well down into the flask. The 
iodine is made up with cold boiled-out distilled water. After each 
estimation the flask is cleaned with a small quantity of concentrated 
hydrochloric acid and washed several times with water. 
Contrasted with the three more recent methods of Gardner and MacLean, 
Stein and Wiseley, and MacLean, the method of Bang has several advantages. 
The quantity of blood required is exceedingly small — only about T c.c. — so 
that several estimations may be made upon the same person in one day 
without inconvenience. Also any method for blood analysis which permits 
of the blood being weighed is to be preferred to one in which it is measured. 
The quantity of blood delivered by an ordinary pipette is very difficult to 
estimate, and is almost certain to differ in different people. As regards the 
method of Stein and Wiseley for measuring blood, the error involved in 
measuring -5 c.c. of blood by dropping into an ordinary graduated centrifuge 
tube may be very considerable. In Bang’s method there is no difficulty 
about filtration. The paper used in soaking up the blood retains all 
the albumen and the reduced copper salt is held in solution. Success- 
ful filtration through an asbestos mat requires considerable experience 
and a very reliable make of asbestos wool. The adding of a known 
amount of glucose to the copper solution in order to obtain a filter- 
able amount of cuprous oxide in the Bertrand estimation makes the 
process easier but no more exact, for any error that occurs affects 
the blood sugar alone. Glucose is, moreover, a difficult substance to 
standardise. The new method of MacLean has the obvious advantage 
over his earlier method that no filtration of cuprous oxide is required, 
but compared with the method of Ivar Bang the quantity of blood used 
is large. 
The outstanding advantage which the Michaelis-Bertrand methods have 
over that of Bang is that the final titration is performed with permanganate, 
whereby an outside indicator is avoided and a very clean end point obtained. 
Still, with a little practice, the exact point at which the green-blue of the 
copper solution gives place to the clear blue of the starch-iodine compound 
can be readily enough determined. 
