36 
SELECTED ARTICLES. 
potash is known to contain, in minute proportions, I did not 
consider it necessary to separate, as it would have occupied 
much more time, and after all would have been attended with 
little practical advantage. My design was, not so much to 
present complete chemical analyses, as to show, in the plainest 
manner, the comparative value of the specimens which I ex- 
amined. The following is the process which I adopted: 
A. Five hundred grains of the specimen under examina- 
tion were dissolved in six or eight ounces of water, heated to 
about 200° Fahr. and the solution filtered. The matter on the 
filter was again washed with a small quantity of water, and 
after being perfectly dried, was weighed, and the amount set 
down as insoluble residuum. 
B. The insoluble matter was treated with dilute nitric acid, 
and then tested with various reagents. It generally consisted 
of carbonate of lime, oxide of iron and silex, although the 
proportions were very various. These I did not think it ne- 
cessary, in each case to determine. 
C. The weight of the filtered solution A, was now deter- 
mined, and to a known portion of it, acetic acid was added 
until the alkali was completely neutralized. 
D. To this neutral solution, solution of acetate of baryta 
was added as long as it caused any precipitation, and the 
whole then filtered ; the resulting sulphate of baryta, being 
dried and weighed, gave by estimation, the amount of sulphu- 
ric acid contained in the whole solution (A.) Supposing this 
acid to be combined with potash, the amount of sulphate of 
potash, in the sample under examination, was easily settled. 
No account was taken of the small quantity of sulphate of 
lime, which might be contained in the solution. 
E. To the filtered solution D, nitrate of silver was now ad- 
ded, as long as any thing was thrown down by it. The re- 
sulting precipitate was dried and weighed. This was chloride 
of silver, and from its weight, by estimation, the whole 
amount of chlorine was determined. Supposing this to have 
been originally combined with potassium, the amount of 
chloride of potassium was deduced. 
