302 Effects of Heat jnodified ly Compression. 



and the tube, a set of fragments of chalk, of a size beyond 

 what could possibly fall between the cylindrical part and the 

 tube, and pressed them down with a blunt tool, by which the 

 chalk being at the same time crushed and rammed into the 

 angle, was forced into a mass of some solidity, which effectu- 

 ally prevented any thing from passing between the large mass 

 of chalk and the tube. In practice, I have found this method 

 always to answer, when done w ith care. I covered the chalk, 

 thus ranuned, with a stratum of pounded flint (o), and that 

 again with pounded chalk (/)) firmly rammed. In this manner 

 I filled the whole of the large tube ^vith alternate layers of silex 

 and chalk ; the muzzle being always occupied with chalk, 

 which was easily pressed into a mass of tolerable firmness, 

 and, suffering no change in very low heats, excluded, the 

 fusible metal in the first stages of the experiment. 



The large tube, thus filled, was placed in the cradle, 

 sometimes with the muzzle upwards, and sometimes the 

 reverse. I have frequently altered my views as to that part 

 of the arrangernent, each mode possessing peculiar advan- 

 tatres and disadvantages. With the muzzle upwards, (as 

 shown in fig. 24. and 25.) the best security is afforded against 

 the intrusion of the fusible metal ; because the air, quitting 

 the air-tube in the working position, occupies the upper 

 part of the barrel ; and the fusible metal stands as a liquid 

 (at a, fig. 2.').) belovi' the muzzle of the tube, so that all 

 communication is cut off between the liquid metal and the 

 inside of the lube. On the other hand, by this arrange- 

 ment, the small tube, which is the fiducial part of the ap- 

 paratus, is placed'at a considerable distance from the breech 

 of the barrel, so as either to undergo less heat than the upper 

 part, or to render it necessary that the barrel be thrust high 

 into the muffle. 



With the muzzle of the large tube downwards, the inner 

 tube is placed (as shown in fig. 22.) so as still to have its 

 muzzle upwards, and in contact with the breech of the 

 laro-c tube. This has the advantage of placing the small 

 tube near to the breech of the barrel : and though there is 

 here less security against the intrusion of liquid metal, I 

 have found that a point of little consequence ; since, when 



the 



