Manufacture of Gilt Buttons . 288 
if the space between A and the iron plate C w as covered 
up with a glass window coming down so low as only to 
leave sufficient room for moving the pan backwards and 
forwards with facility. If the sides were also glass in- 
stead of brick-work it would be still better, as the work- 
people would be able to have a full view of their work 
without being exposed to the fumes of the mercury* 
which, when volatilised by heat communicated to the 
pan by the heated iron plate over the fire-place, would 
ascend into the top A, appropriated for its reception, and 
descend into the tub G, covered at top and filled pretty 
high with water. By this means the hearth would, in 
fact, become a distilling apparatus for condensing and re- 
covering the volatilised mercury. In the tub Gr the prin- 
cipal part would be recovered ; for, of what may still 
pass on, a part would be condensed in ascending the tube 
II, and fall back, while the remainder would be effec- 
tually caught in the tub or cask I, open at the top and 
partly filled with water. The latter tub should be on the 
outside of the building, and the descending branch of 
the tube H should go down into it at least 18 inches, but 
not into the water. The chimney or the ash-pit should 
be furnished with a damper to regulate the heat of the 
rfie. 
The water may be occasionally drawn out of the tubs 
by a siphon, and the mercury clogged with heterogene- 
ous matter may be triturated in a piece of flannel till it 
passes through, or placed in a pan of sheet iron, like a 
dripping-pan, in a sufficient degree of heat, giving it a 
tolerable inclination, so that the mercury, as it gets warm* 
may run down and unite in the lower part of the pan. 
But the mercury will be most effectually recovered by 
exposing the residuum left in the flannel bag to distilla- 
tion in a retort made of iron or of earthenware. 
When the mercury is volatilised from, the buttons, or* 
VOL. I, G g 
