2 43 
than at any other period ; and that when cold, and In pro- 
portion to the absence of heat, so will the volume of the 
metal be diminished. 
1st, Then, Shrinkage appears to be dependent upon two 
causes ; the gravitation of the fluid metal, and the expan- 
sion of the mould. The latter, I conceive, acts a very pow- 
erful part : the immense quantity of caloric combined 
with the iron is in part easily and almost instantaneously 
communicated through the sand to the iron box : this 
creates a disposition to expand, in which it is greatly as- 
sisted by the great pressure of fluid iron. That portion 
of the metal in contact with the interior of the mould is 
the first to lose its fluidity, and is acted upon and forced 
to give way in the same ratio of expansion before the sub- 
tle and denser fluid. The diameter of the shell of the gun, 
is at this period increased in every part ; the fluid iron in 
the interior descends to occupy the enlarged space, and the 
head of the gun presents an increasing chasm like the con- 
cave of a sand glass. In proportion as the cast iron re- 
solves itself into a solid, a diminution of pressure should 
take place upon the mould : this would inevitably follow, 
were not its force replaced by tire increased volume of the 
metal passing into a solid state, which is equivalent to that 
law which I have termed 
2d, Expansion . Of the extent of this operation we 
may judge from the following facts : — - All patterns of cast- 
ings are made somewhat larger than the piece of goods is 
wished to be : in common cases l-.8th of an inch to the 
foot is allowed, but in many cases the allowance will be 
nearly 3-16ths of an inch. In the case of the gun, there- 
fore, the mould would be plus the allowance upon the pat- 
tern what space was gained by beating the pattern to loose 
it from the sand, and all the extra space acquired by the 
increased volume of the consolidating iron. These, taken 
collectively, may amount to I -4th or 5-10ths of an inch ; 
