The author next referred to the great excellence for chilling purposes possessed by some American pig irons, and to the fact that iron of a given carbon content derived from some ores and fluxes differed much in chilling properties from iron holding a similar proportion of carbon--free and combined--derived from other ores and materials. Those irons are best which develop the hardest possible chill most uniformly to the desired depth without producing a too abrupt line of division between the hard white skin and the softer gray body. A medium shading off both ways is wanted here, as in all things. The impossibility of securing a uniform quality and chemical composition in any number grade of any brand of pig iron over a lengthened period was adverted to. Consequent from this a too resolute faith in any particular make of pig iron is likely to be at times ill-requited. Occasional physical tests, accompanied with chemical analysis of irons used for chilling, were advocated; and the author was of opinion it would be well whenever a chilled casting had enjoyed a good reputation for standing up to its work, that when it was retired from work some portions of it should be chemically analyzed so as to obtain clews to compositions of excellence. Some of the physical characteristics of chilled iron, as well as the surprising locomotive properties of carbon present in heated iron, were noticed. Attention was called to some German data, published by Dr. Percy in 1864, concerning an iron which before melting weighed--approximately--448¼ lb. per cubic foot, and contained--approximately--4 per cent. of carbon--3¼ being graphitic and ¾ combined. The chilled portion of a casting from this had a specific gravity equivalent to 471 lb. per cubic foot, and contained 5 per cent. of carbon, all combined. The soft portion of the same casting weighed 447¾ lb. per cubic foot, and contained 34.5 per cent. of carbon--31.5 being graphitic and 3.5 combined. Mr. Morgans doubted whether so great an increase in density often arises from chilling. Tool steel, when hardened by being chilled in cold water, does not become condensed, but slightly expanded from its bulk when annealed and soft. Here an increase of hardness is accompanied by a decrease of density. The gradual development of a network of cracks over the face of a chilled anvil orbit while being used in tilt hammers was mentioned. Such minute cleavages became more marked as the chill is worn down by work and from grinding. Traces of the same occurrence are observable over the surface of much worn chilled rolls used in sheet mills. In such cases the sheets get a faint diaper pattern impressed upon them. The opening of crack spaces points to lateral shrinkage of the portions of chilled material they surround, and to some release from a state of involuntary tension. If this action is accompanied by some actual densification of the fissured chill, then we