Brass, 103 
A bell-metal was first made with 80 parts of copper and 
20 of tin, both pure. 
Of this alloy 100 parts were put on a muffle and expos- 
ed for a certain time to a red heat, with access of air. The 
whole was converted to a grey oxyd and weighed 104 
parts. It was then heated strongly for half an hour, and 
yielded a button of pure copper weighing 54 parts. Con~ 
sequently 26 parts of copper remained unreduced and 
mixed with the scorias of the oxyd of tin. 
A hundred parts were calcined as above till they in- 
creased to 117. This being strongly heated for half an 
hour gave a brown mass, from which no copper separat- 
ed except a few interspersed globules. This therefore 
was too much oxygenated, that is, the tin had combined 
with so much oxygen in the first operation that it could 
not afterwards separate any from the oxyd of copper. 
A hundred parts were treated in the same way till they 
increased to 112. This melted into a brown mass con- 
taining more globules of copper than the last, but still the 
greater part remained unreduced. 
Hence it appears that even an increase of 12 on 100 is 
too much, and by other experiments, the above chemist 
fixes the proper point of oxydation at which oxyd of cop- 
per is reducible by oxyd of tin to be from 5 to 7 on 100 
of common bell-metab However, when too highly oxy- 
genated metal is mixed with the requisite proportion of 
fresh metal, the former serves as a reducing flux to the 
latter, and the whole may then be made to yield a large 
portion of purified copper. This very ingenious plan has 
been found to answer extremely well in experiments 
made in the large way, as will presently be mentioned. 
Several substances have been tried that might assist in 
the first oxygenation of the bell-metal and shorten the pro* 
cess. Of these, nitre and oxyd of manganese answer the 
most completely. For experiment in the small way, ZOO 
