ADULTERATION OF POTASSA BY SODA. 213 
perty of soda, of being able to absorb a very great quantity 
of water of crystallization — a property which potassa does 
not possess — presents to us a means of determining, in an 
extremely simple manner, although only approximative^, 
the proportion of soda contained in potassa. 
With respect to the accuracy of this process, it doubtless 
still leaves much to be desired ; for I am convinced, from 
a great number of investigations, that it may yet give rise 
to an error of two or three per cent., even when we have 
become familiarized with its execution by considerable 
practice and that it is a pure chance if the error remains 
less than a hundredth. Moreover, the impurities which 
are ordinarily found in potassa do not interfere much with 
its accuracy ; these are ordinarily substances, such, for ex- 
ample, as the sulphate of potassa, chloride of sodium, chlo- 
ride of potassium, silica, &c, which do not take up water of 
crystallization, and the quantity of water necessary merely 
to moisten those ordinary impurities of potassa is so small, 
that it is not necessary to take account of it. 
Moreover, it is necessary, in the examination of a sample 
of potassa by this process, to commence by freeing it from 
the water which it contains, by putting it in an oven, or 
any other warm place — or, better still, to heat it to redness, 
in order to convert into simple carbonates the sesquicarbon- 
ates which it may contain. 
Although, according to these considerations, this process 
cannot be sufficient for the chemist, and should never be 
adopted as a basis in important cases, it may, however, very 
well serve the purpose of persons unacquainted with che- 
mistry, (ordinary merchants and manufacturers,) on account 
of its great facility of execution. 
Moreover, the circumstance already mentioned — that, on 
the one hand, this process cannot pretend to accuracy, and 
that, on the other hand, it requires a long time for its exe- 
cution, (which appears to be an inconvenience for ordinary 
use, which requires the utmost speed) — this circumstance 
19* 
