ferent states of shrinkage and expansion, as observed m 
east iron* 
In doing this I shall divide shrinkage into two distinct 
operations r 1st, Shrinkage, properly so called, when a 
mass of iron diminishes or shrinks within itself, and would 
actually displace a smaller quantity of water, and when no 
degree of heat short of fusion would make it occupy its 
former bulk or volume* 2d, Contraction, or that diminu- 
tion of superficial measurement which any body under- 
goes by evolving its caloric* The surface in this case is 
never injured ; the casting will be found less than the pat- : 
tern from which it was formed, and simple heating will 
restore it to its greatest original volume* 
The former of these properties cannot exist without the ; 
latter, but this last may take effect in full force in many 
minor operations without any appearance of shrinkage. 1 
only say appearance ; for I believe, abstractly speaking, 
the one never takes place without the other, though in such 
various minute degrees that it is often difficult to form any 
estimate of the quantity. 
In casting pieces of ordnance we are enabled to judge of 
the conjoint effects of shrinkage, contraction, and expan- 
sion. We shall suppose that a gun mould of any given 
length is filled with fluid cast iron not subject to these 
laws ; then the size and shape of the gun r when cold, 
would exactly correspond to the dimensions of the mould* 
but finding that the piece of casting was considerably al- 
tered, that it had shrunk interiorly, was diminished in point 
of lengthy and had lessened its diameter, we must seek for 
a solution of these facts in the explanation of the causes 
respectively. 
First assuming, what shall be hereafter proved by direct 
experiment, that cast iron occupies less volume when fluid 
than when solid ; that in the act of the arrangement of the 
molecule towards consolidation, it occupies a larger bulk 
