102 EFFECTS of HEAT 
contributed very much both to the convenience and fafety 
of thefe experiments. _ 
In this view, making ufe of the brick-furnace with the 
vertical muffle, already defcribed in page 93. I ordered a pit 
(aaa, fig. 20.) to be excavated under it, for the purpofe of 
receiving a water-veflel. This veflel (reprefented feparately, 
fig. 21.) was made of caft iron; it was three inches in diame- 
ter, and three feet deep; and had a pipe (d e) ftriking off from 
it at right angles, four or five inches below its rim, communi- 
cating with a cup (ef) at the diftance of about two feet. 
The main veflel beng placed in the pit (a4) directly below the 
vertical muffle, and the cup ftanding clear of the furnace, wa- 
ter poured into the cup flowed into the veffel, and could thus 
conveniently be made to ftand at any level. “(The whole ar- 
rangement is reprefented in fig. 20.) The muzzle of the barrel 
(g) being plunged into the water, and its breech (+) reaching 
up into the muffle, as far as was found convenient, its pofition 
was fecured by an iron chain (gf). The heat communi- 
cated downwards generally kept the furface of the water 
(at c) in a ftate of ebullition; the wafte thus occafioned being 
fupplied by means of the cup, into which, if neceflary, a con- 
ftant ftream could be made to flow. | 
As formerly, I rammed the carbonate into a tube of porce- 
lain, and placed it in a cradle of iron, along with an air-tube 
and a pyrometer; the cradle being fixed to a rod of iron, 
which rod I now judged proper to make as large as the barrel 
would admit, in order to exclude as much of the fufible me- 
tal as poflible ; for the expanfion of the liquid metal being in 
proportion to the quantity heated, the more that quantity 
could be reduced, the lefs rifk there was of deftroying the 
barrels. 
In the courfe of practice, a fimple mode occurred of remo- 
ving the metal and withdrawing the cradle- it confifted in pla- - 
cing 
