188 
Proceedings of the Royal Society of Edinburgh. [Sess. 
N . . 
performed. If n he the number of c.c. of ^qq iodine used in the titration, 
the amount of glucose is calculated from the formula 
71 C.C. — *16 c.c. , . 
^ — glucose m milligrams. 
By introducing certain modifications into Bang’s method we have been 
able to secure very satisfactory results. Our process is as follows : — 
Instead of using a torsion balance the filter paper is weighed in a small 
weighing bottle with a well-fitting stopper, the blood is then soaked up 
into it, care being taken to leave the margins fairly clean, and the paper 
replaced in the weighing bottle and weighed again. A small balance which 
can weigh accurately to milligrams is used, and the time occupied in 
weighing the blood is very short. Loss due to evaporation is almost 
completely avoided, and the time required for the frequent determination 
of the constants of a torsion balance is saved. The inaccuracy of a torsion 
balance, unless frequently calibrated, is well known to all physicists. The 
solution used for coagulating the albumen is boiled for some minutes to 
render it air-free, a small quantity of distilled water being added to allow 
for evaporation. A few minutes — about five — should be allowed for the 
blood to dry before it is coagulated, and it is weR not to shake the tube 
after the boiling solution has been poured in. Usually the albumen is 
completely retained in the paper, but should coagulation be incomplete 
the specimen may be rejected at once, as estimations of glucose in solutions 
containing traces of albumen are most inaccurate. An hour is allowed for 
the glucose to diffuse into the surrounding solution, which is then transferred 
to the small flask, the paper washed with a further quantity of potassium 
chloride solution, which has been boiled to render it air-free and cooled, 
and the washings added to the contents of the flask. Neither in Bang’s 
original papers, nor in any account of his method which we have seen, is 
mention made of having the coagulating solution air-free. If no precaution 
be taken to attain this, the method amounts to adding 1 c.c. of an air-free 
solution to 13 c.c. of an air-containing one. The copper solution is also 
boiled shortly before it is to be used, to render it air-free. A small 
quantity of distilled water is added to a few c.c. of the solution and the 
whole boiled for some minutes and cooled. 1 c.c. is then added to the 
glucose solution in the small flask. Experiments with pure glucose led to 
the belief that the most accurate results are obtained with the flame 
so adjusted that the contents of the flask begin to boil in from 1 minute 
15 seconds to 1 minute 25 seconds. The efficient sealing of the flask is 
