82 EFFECTS of HEAT 
plete compreflion: and, at the fame time, it illuftrates what 
has happened occafionally in nature, where the bituminous 
matter feems to have been driven by fuperior local heat, from 
one part of a coaly bed, though retained in others, under the 
fame compreflion. The bitumen fo driven off being found, in 
other cafes, to pervade and tinge beds of flate and of fandftone. 
I was employed in this purfuit in {pring 1800, when an 
event of importance interrupted my experiments for about a 
year. But I refumed them in March 1801, with many new 
plans of execution, and with confiderable addition to my ap- 
paratus. 
In the courfe of my firft trials, the following mode of execu- 
tion had occurred to me, which I now began to put in pra¢tice. 
It is well known to chemifts, that a certain compofition of differ- 
ent metals *, produces a fubftance fo fufible, as to melt in the 
heat of boiling-water. I conceived that great advantage, both 
in point of accuracy and difpatch, might be gained in thefe ex- 
periments, by fubftituting this metal for the baked clay above 
mentioned: That after introducing the carbonate into the 
breech of the barrel, the fufible metal, in a liquid ftate, 
might be poured in, fo as to fill the barrel to its brim: 
That when the metal had cooled and become folid, the breech 
might, as before, be introduced into a muffle, and expofed 
to any required heat, while the muzzle was carefully kept cold. 
In this manner, no part of the fufible metal being melted, but 
what lay at the breech, the reft, continuing im a folid ftate, 
would effeGtually confine the carbonic acid : That after the ac- 
tion of ftrong heat had ceafed, and after all had been allowed 
to cool completely, the fufible metal might be removed entire- 
ly from the barrel, by means of a heat little above that of boil- 
ing water, and far too low to occafion any decompofition of 
the 
* Eight parts of bifmuth, five of lead, and three of tin. 
