Correspondence. 65 
The treatment described under (b) would decompose a silicate, 
while the method under (a) would not. As the two series of results 
are fairly concordant, the conclusion is that the silica exists as a fine 
sand disseminated through the rock. A private communication from W. 
H. Norton, of the Cornell College department of geology, states that 
he came to the same conclusion while studying the rock with a petrolog- 
ical microscope. The method described under (a) is simpler than (b), 
and the work can be done in a much shorter tim.e. It is therefore to be 
preferred in the analysis of rock of this kind. 
2. The condition of the iron. 
A gram of the substance was introduced into a flask of I20cc. capa- 
city, fitted with a bulb tube and Bunsen valve to prevent oxidation of 
the iron. It was dissolved in dilute hydrochloric acid. A few drops of 
the cooled solution were then withdrawn with a capillary tube and tested 
with a crystal of potassium ferricyanide. No suggestion of a blue color 
resulted, showing the iron to be in the ferric condition. This increases 
the value of the rocks as a building material, as ferrous carbonate is an 
unstable substance with a tendency to change to the ferric condition. 
A complete analysis of the specimen resulted as follows : 
CaCOa 53.62^ 
MgCO-^ 44.96;^ 
SiO. 0.83?^ 
ALO3 0.25,-^ 
Fe.O. 0.34^ 
Total 100.00;^ 
The specimen is nearly a true dolomyte which contains : 54.35 per 
cent calcium carbonate, 45.65 per cent magnesium carbonate. 
This method of analysis was employed : After removing the silica 
according to (a), a gram or two of pure ammonium chloride was added 
to the filtrate to prevent the precipitation of magnesium. It was then 
heated to boiling and a small excess of ammonia added which precipi- 
tates iron and alumina. They are determined together and then dis- 
solved in the crucible with warm hydrochloric acid. The solution is 
treated with caustic potash which precipitates the iron and dissolves the 
alumina. The iron is filtered ofif and discarded because it can not be 
thoroughly washed from the caustic potash. The filtrate is slightly acid- 
ified with hydrochloric acid and the aluminum is precipitated with 
freshly prepared ammonium sulphide. The precipitate when heated in a 
crucible becomes ALOs. The filtrate from the iron and alumina, con- 
taining calcium and magnesium, is heated to boiling and precipitated 
with an solution of ammonium oxalate, care being used to avoid 
much excess of the reagent. The precipitate was allowed to stand 
eight to twelve hours before filtering. The well washed precipitate of 
calcium oxalate containing also a small quantity of magnesium oxalate 
was dissolved in warm dilute hydrochloric acid and the solution was 
made alkaline with ammonia. This precipitates the calcium oxalate and 
leaves the magnesium in solution. This, with the main portion of the 
