129 



noAV exist only in the appendage to it, whith is a super- 

 numerary part, to be afterwards separated from it. This 

 appendage ought to be of the form of a prism, and as nearly 

 that of a cube, as the operation of moulding it in the sand 

 will permit; for, in this gross shape, the metal in it will be 

 the longer cooling. It should be connected with that part 

 of the mirror, which is uppermost in the flas4c, and joineA 

 to it by a neck, equal in thickness to the edge of the mirror, 

 (but so posited, that the face of the mirror may project a 

 little above it,) and, in breadth, about twice the thickness. 

 This neck ought to be as short as possible, i. e. just so as 

 to permit it to be nicked round with the edge of a file, in 

 order to break off the prism from the iiiirror when cast: 

 for thus the heat of the large contiguous body of the prism^ 

 will keep the neck from congealing ; which, if it happened, 

 would stop the liquefied metal, in the prism, from running 

 down into the mirror. And, to prevent this, the prism ought 

 not to form directly a part of the main jet or ingate, by 

 which the metal is poured into the flask; for so the jet 

 would cool sooner than the large mass of the mirror, and 

 bear off the weight of the atmosphere, which ought to press 

 on the fluid metal in the prism underneath, and force it 

 down into the mirror, to fill up all vacuities in it. Both 

 the prism and the mirror, therefore, ought to be filled by 

 a lateral channel, opening (from the principal ingate) into 

 ,the top of the prism ; which latter should be formed broad 

 land flat, and not taper upward, like a pyramid, lest, by 

 cooling where it grows narrow, it might form a solid arch, 



and 



