Correspondence. 59 
platinum crucible with seven grams of anhydrous sodium carbonate. 
The sodium carbonate used was a good quality of Merck's manu- 
facture. It was further purified by dissolving a quantity in water 
and filtering. This after fractional crystallization seemed to be 
entirely free from silica. The covered crucible was heated for 
fifteen minutes with a Bunsen flame, then with a blast lamp to 
complete fusion. The cooled mass was transferred to a porcelain 
evaporating dish and dissolved in dilute hydrochloric acid. Evap- 
oration was continued on the water bath until crystals began to 
appear. It was then stirred with a glass rod until a fine dry 
powder was obtained. It was treated with hydrochloric acid as in 
method 2 before described. The insoluble residue obtained by this 
method amounted to 0.53%. This treated with sulphuric and hydro- 
fluoric acids left a residue of 0.16%. The results are fairly con- 
cordant with those obtained by methods 1 and 2, but on account 
of the difficulty of getting pure sodium carbonate, and the length 
of time and the labor necessary to obtain the result, this process 
is not to be commended with rocks of this character. Indeed 
method 1 is greatly to be preferred, whenever circumstances per- 
mit, on account of its simplicity, and the shortness of time required. 
During the course of this work our attention was called to the 
estimation of silica as outlined by Treadwell* in his excellent 
treatise on quantitative analysis. After removing the insoluble 
residue according to method 2, it is stated that "as much as 5 per 
cent of the total amount (of silica) may remain in the filtrate. In 
order to remove this the filtrate from the first precipitate is once 
more evaporated to dryness on the water bath and kept on the 
hot water bath for one or two hours or more." It is then filtered 
after treatment with hydrochloric acid and water in the usual 
manner. 
Several determinations were made in which the suggestions of 
Treadwell were strictly carried out. One gram, five grams and ten 
grams of substance were used. In no case was even a trace of 
residue obtained by the second treatment. The experiment was 
varied by using a specimen of argillaceous limestone that contained 
18 per cent of silica. A second portion of residue could not be 
obtained from this specimen. 
A complete analysis of the Bedford limestone resulted as fol- 
lows: 
Ca Co 3 93.55% 
Mg Co 3 5.42% 
Si O 2 0.55% 
Fe 2 O 3 and AF O 3 0.50% 
100.02% 
The condition of the iron was tested by placing three grams 
of the powdered rock in a flask of 12 c c capacity, fitted with a bulb 
* Treadwell-Hall's Quantitative analysis, page 384. 
