COMMERCIAL POTASH OF THE STATE OF NEW YORK. 43 
added to a given weight of the sample, (previously dissolved 
in water, and the solution filtered,) until by a test paper, it is 
shown, that the alkali is exactly neutralized. The amount of 
acid required to produce this effect, will, if its strength be 
exactly known, indicate the proportion of pure alkali, con- 
tained in the specimen under examination. 
It is evident, however, that this process is insufficient to 
detect the presence of soda, and when that substance is in 
large proportion, it is too objectionable to be relied on, for the 
reason that the saturating powers of potash and soda are very 
different. For example, fifty parts of soda will saturate as 
much of any given acid as seventy -five parts of potash; so 
that in practising with this test, a mixture of twenty-five parts 
of soda and 37.5 parts of potash =62.5, would give the same 
result, as seventy-five parts of potash. In this instance then, 
there would be an error of 12.5 per cent., to say nothing of 
the difference in the value of soda and potash. 
4. If the above remarks are correct, it follows as a conse- 
quence, that in all cases where carbonate of soda is mixed with 
potash, the relative proportions of these alkalies must be de- 
termined, before we can fix the real value of the sample under 
examination. 
The only unexceptionable mode of separating potash from 
soda, is by means of nitro-muriate of platina; but this is an ex- 
pensive article, and its successful employment requires a 
nicety of manipulation which can be expected only from the 
experienced chemist. An approximation to a correct decision 
on this point may be attained by adopting the steps prescribed 
in D and C, in the preceding formula. By adding to a known 
quantity of the alkaline solution, an amount of nitric acid suf- 
ficient for its neutralization, the sulphates and chlorides may 
be thrown down, first by the acetate of baryta, and afterwards 
by the nitrate of silver. If the latter test occasions an abun- 
dant precipitate, we may infer that the solution contains an 
admixture of common salt; for although chloride of potassium 
is contained in the purest samples of potash, it seldom con- 
stitutes more than two or three per cent. 
