104 
Brass* 
parts of bell-metal were powdered atid mixed with about 
14 of nitre, and heated, at first slowly, afterwards strong- 
ly* and gave from 57 to 63 parts of copper, very mallea- 
ble, but not quite pure. With a larger quantity of nitre 
the copper is purer* but more is lost in the scoriae. 
A hundred parts of bell-metal mixed with 25 parts of 
oxyd of manganese, covered with broken glass, and 
strongly heated for an hour, gave 36 parts of very good 
copper. It is of some importance to add glass or some 
other saline flux, as this brings the whole into thorough 
fusion, and allows the copper to subside through the vitri- 
fied oxyd of manganese. Where this is not done (and in 
the large way it would add to the expense) much of the 
copper is entangled in the vitrified manganese, and a sub- 
sequent operation is required to separate it.* 
In the processes of two other chemists, salt and sand 
have each been found useful additions in the refining of 
bell- metal. 
We may add the detail of two trials made in the large 
way by Pelletier and Darcet by order of the French go- 
vernment, the one in which a portion was first oxy dated* 
and this used as a reducing flux for the remainder, with- 
out any foreign addition whatsoever : and the other in 
which the refining was assisted by oxyd of manganese* 
It may be premised that the manipulation of large quanti- 
ties is not the same as with small, particularly in the case 
<tf reduction to gross powder, which may be easily done in 
an iron mortar on a few ounces or pounds, but not with any 
economy on large weights. Large masses therefore are 
first heated red-hot, when they may be easily broken up 
and spread about by an iron bar, and this is the way con- 
stantly practised. 
The first process was that of refining without addition* 
For this purpose 400 pounds of metal were put on a fur- 
nace, and when red-hot broken up with an iron bar, ani 
* All this strongly bears upon the refining of Copper. 
