166 C. J. WHITE. 
Silica.—In place of a single evaporation followed by 
heating to 120° the double evaporation recommended by 
Hillebrand was employed, the second was found to account 
for about ‘2% silica. The purity was tested by hydrofluoric 
acid and the smalJl quantity of Fe etc. carried down was 
treated in the usual way. 
Ferrous Iron— Here the method devised by Cooke’ was 
employed, i.e., the finely powdered substance was heated 
with HF and H.SO; on a water bath and in an atmosphere 
of carbon dioxide, then cooled and titrated with potassium 
permanganate. 
Manganese.—The presence of manganese was indicated 
by the colour of the mass after fusion, but owing to the 
delicacy of this test and also to the fact that sulphuretted 
hydrogen failed to bring down any precipitate in 20 minutes 
(volume of solution = 500 cc.) manganese was thought to 
be present only as a trace. Consequently in the first 
analysis (rejected) its estimation was neglected. However 
the discolouration of lime combined with a shortage of °5% 
in the total, caused special precautions to be taken in the 
two succeeding analyses. Of all the methods tried for the 
irou Manganese separation the most satisfactory was the 
ordinary one by ammonia (three precipitations). The 
manganese was precipitated by evaporating the solution to 
less than 200 cc., passing sulphuretted hydrogen and then 
allowing to stand ina corked flask over night; it was eventu- 
ally weighed as Mn,O,. The presence of alumina with 
such small quantities of iron and manganese seemed to 
vitiate the basic acetate method and the barium carbonate 
method of Treadwell* was found to be cumbersome. A 
colorimetric method recently proposed* was found to give 
’ Hillebrand, p. 52. 
* Treadwell (Hall)—Quantitative Analysis, p. 121. 
* J. G. Cooke—Amer. Journ. Sci., (2) xLIv., p. 347, 1867. 
* Nazareno Tarugi—Gazzetta, 1906, p. 332. Abstract in Journ. Chem, 
Soc., 1906, p. 631. 
