216 ADULTERATION OF POTASSA BY SODA. 
then divided into 25 equal parts; then I continued to gra- 
duate the tube towards the upper part, until I had marked 
100 parts and I added the numbers, ascending from 6 to 5, 
from the lower parts towards the upper. I afterwards broke 
the tube at about two inches above the place marking 100, 
and I united with the lamp the acute angles. 
This tube, thus prepared, was tried in several ways, un- 
til I was perfectly convinced that each part exactly answered 
to one grain of pure carbonate of potassa. 
This tube then served as a standard instrument for testing 
a suspected sample of potassa : for this purpose we proceed 
in the manner I am about to describe. 
Two portions, of 100 grains each, of the potassa to be 
examined, are weighed out. The first 100 grains are tested 
by the ordinary process of alkalimetry ; the quantity of test- 
acid employed in the Table, No. L, is ascertained, and the 
corresponding quantity of carbonate of potassa is noted; 
then the second 100 grains are dissolved in 8 or 10 times 
their weight of the solution of bitartrate of potassa, com- 
pletely saturated at the ordinary temperature, and of which 
a very large quantity should always be kept in reserve. 
The solution is filtered in a flask capable of containing 
about one pound of water, and the filter is washed with the 
saturated solution of bitartrate of potassa. Then a quantity 
of tartaric acid, in fine powder, is added, equal to that indi- 
cated by the adjoining Table, (No. II.,) that is to say, the 
quantity necessary to convert into bitartrate all the alkali 
which is found in the potassa (potassa or soda, or both at 
once, carbonated or caustic.) This done, a sufficient quan- 
tity of the same solution of bitartrate of potassa at a known 
temperature is poured into the flask, so that its whole 
quantity may amount to about 6,400 grains. A slight ex- 
cess or deficiency is generally without inconvenience. In 
order not to be obliged to weigh every time, the height ob- 
tained by the liquor may be marked with a file. 
The flask is then corked, put in a warm place, and care- 
