THE DETERMINATION OP FERROUS IRON. 
BY NICHOLAS KNIGHT. 
Robert Mauzelius, in the Geological Survey of Sweden (Sveriges Geol. Under- 
seekning), calls attention to a universal defect in the estimation of ferrous iron 
in minerals and rocks. He finds that a considerable error arises from grinding 
the rock in the ordinary way to a fine powder, owing to the oxidation of ferrous 
to ferric iron. If the specimen is reduced to a coarse powder instead of being 
finely pulverized, a more accurate result is attainable. This source of error 
has hitherto escaped >the attention of analysts. 
We have investigated specimens of siderite with a view to determine the 
oxidation effect when the mineral is finely pulverized. In each case only the 
ferric oxide was estimated, and the method of Berzelius as modified and im- 
proved by Bunsen was employed. 
About a grm. of the specimen was weighed each time in a 200 cc. flask which 
was fitted with a bulb tube and Bunsen valve. The substance was dissolved 
in a small quantity of hydrochloric acid, and a crystal of sodium carbonate was 
introduced to make an atmosphere of carbon dioxide. The flask was cooled as 
quickly as possible with water, and the ferric iron was precipitated with barium 
carbonate. The precipitated iron was at once filtered with suction and washed 
twenty or thirty times with cold water. The ferric carbonate was dissolved in 
hydrochloric acid, the excess of barium removed with sulphuric acid, and the 
iron precipitated with ammonia. The ores contained only about 1% per cent 
of residue insoluble in hydrochloric acid. Two different specimens of the min- 
eral were studied. 
1. The specimen was run through a Taylor hand ore-crusher, and came out 
as a coarse powder. The mean of several fairly concordant results was 2.68 
per cent Fe203. The coarse powder was next so finely ground in an agate mor- 
tar as not to feel gritty when placed between the teeth. It gave 3.26 per cent 
FeiOs. 
2. This specimen was enveloped in heavy paper, and reduced to a coarse 
powder in an iron mortar; 2.67 per cent of Fe203 was obtained. This coarse 
powder was finely ground in an agate mortar, when 3.11 of Fe203 per cent was 
obtained. 
The oxidation is due to the heat caused by friction, and possibly in a lesser 
degree to the greater surface exposed to the air when the mineral is in a very 
finely divided state. The coarse powder is easily acted upon by the hydro- 
chloric acid, and there is no necessity for making the determination with a very 
fine powder. ^ 
Our thanks are due to Mr. Robert H. Lott for making the analyses herein 
described. 
( 105 ) 
