ON THE REDUCTION OF IRON BY HYDROGEN. 13 
long is firmly screwed. The open end of the tube is closed 
by the hollow iron stopper h. 
4th. The furnace consists of an oblong enclosure of bricks 
without mortar. After the first two layers a number of iron 
rods are placed transversely to support the fire, then two or 
three layers on this making the interior 9 inches wide, 36 
inches long, and 6 inches deep to the grating. 
5th. The arrangement e consists of a common half-gallon 
glass jar connected with the redaction tube on one side, 
and from the other a bent tube issues with a bulb near the 
extremity, the open end dipping in water. Having pre- 
pared the apparatus, the next step is to choose the form of 
of oxide of iron to be treated. MM. Soubeiran and Du- 
blanc employed the anhydrous sesquioxide, and the sub- 
oxide obtained by calcining the oxalate. I have used the 
sub-carbonate of iron (precipitated carbonate) of the shops 
with perfect success, and this is the form most readily at- 
tainable. It should be absolutely free from saline matter, as 
sulphate of soda or potassa, because these salts are de-oxi- 
dized by hydrogen, and sulphurets of potassium and sodium 
formed, which even in minute quantity give a sulphurous 
taste to the iron as well as attract moisture. 
Having the apparatus and oxide ready, a very important 
step consists in introducing the latter into the reduction tube 
in such a manner that it will expose a very large surface to 
the action of the gas. Soubeiran and Dublanc used a suc- 
cession of flat shelves of wire gauze kept apart by small iron 
bars. This is inconvenient because of the small diameter 
and great iength of the tube, A better plan is to provide 
pieces of wire gauze seven-eighths of the length of the tube, 
and from three to six inches wide, bent in a semi-cylindri- 
cal shape. The widest is then covered with a layer of sub- 
carbonate about a quarter of an inch thick; on this another 
gauze is placed, and likewise covered with the sub-carbo- 
nate, and gauze after gauze with iron intervening is thus 
built up until the pile assumes a size almost equal to the 
VOL. XIII. — NO. II. 2 
