[ 2& 5 ]

and united with the rest. I tried different kinds of castiron, with the fame event. Even forged iron, thoughconsiderably heavier than cast iron in its solid state, wasfound to be lighter than melted cast iron ; for it floatedon the surface, and when pushed to the bottom, it roseup again, and this repeatedly till it was dissolved by themelted metal. Had the solid pieces barely swum on thetop, it might have been suspected that they were keptfrom sinking only by want of sufficient fluidity in themelted iron : but their constantly rising up from the bot-tom, seems a proof of their being lighter than the fluid.

It appears therefore that melted iron is really of greaterspecific gravity, or more dense, than solid iron, and con-sequently that in fixing or becoming solid, it becomeslighter, or expands into a larger volume; and yet, thatwhen grown cold, it does not press against, or keep dis-tended, the vessel or cavity it was poured into. Nor dothese different effects seem to be at all repugnant to oneanother. It is not pretended that iron expands at anyother period of its cooling, than in the instant of its pas-sage from a fluid to a solid state : after this time it con-tracts like the other metals. The internal cavities areagreeable to this account: the outer surface first expand-ing and fixing, a vacuity would remain under it if thenext did not also expand : a vacuity must necessarily remainat last, which can be filled only by the subsequent contrac-tion ; and its not being filled seems to shew that the expan-sion is greater than the contraction.

To judge in some measure of the degree of the con-traction, I melted some cast iron, and poured it into along narrow iron ingot mould. The ingot proved insome parts convex on the surface, and in others a littledepressed : it was shorter than the mould by nearly threeparts in three hundred and thirty-two, or one part in a

M m hundred.