
1905.] On the Physiological Processes of Green Leaves. 37 
will not be accumulated in the result, but will only affect those results in the 
proportion the errors of measurement bear to the true volume. 
Certain precautions are necessary if the highest possible degree of accuracy 
is required. | 
In order to avoid an undue increase in the volume of the liquid under- 
going titration, which would of course diminish the sensitiveness of the 
reaction, it is advisable to use in the early stages of the first titration an acid 
of a strength varying, according to circumstances, between normal and 
6-normal, until the pink colour of the phenol-phthaléin becomes somewhat 
faint : this is followed by 1/10-normal acid to the complete disappearance of 
the colour. Whilst the stronger acid is being run in it is advisable to keep 
the solution in constant rotation, otherwise acid carbonate is locally decom- 
posed and carbon dioxide evolved, a misfortune which cannot be rectified. 
Caustic soda made from metallic sodium should alone be used, since 
a trace of alumina or iron interferes with the sharpness of the second 
titration, A trace of alkaline silicate also interferes with the delicacy of the 
phenol-phthaléin reaction, a fact which Letts and Blake have also observed,* 
but this is not so important a drawback when the above double-titration 
method is used as it is in methods which are dependent on the exact 
titration of the whole of the alkali at start and finish. 
The actual titrations in our experiments were carried out in cylindrical 
vessels covered with an india-rubber cap furnished with two tubulures 
through which the delivery-tubes of the burettes were passed. The use of 
this cap is unnecessary for the second titration. The burettes used for the 
soda-solution are filled from stock-bottles through side-tubes, and are fur- 
nished at the top with guard-tubes of soda-lime. 
In determining the final point in the first titration the meniscus of the 
liquid is carefully observed, since residual colour can be seen there when it 
is quite inappreciable in the body of the liquid. 
For the second titration a colour-control was always used, containing 
the same ainount of methyl-orange as the liquid titrated, and brought to 
a constant arbitrary tint of acidity. 
When all these precautions are taken the method is sensitive to extremely 
small differences of carbon dioxide content in the soda-solution employed. In 
the paper already cited | we have given a series of experiments made with a 
view to test its limits of delicacy. There is no difficulty in determining in 
100 cc. of an alkaline solution differences in the amounts of carbonate 
* ¢ Dublin Soc. Proe.,’ vol. 9 (1900), p. 152. 
ft ‘Phil. Trans.,’ B (1900), p. 291. 
