to be, though it is not quite clear to me how it can be so if steel is a chemical compound. However this may be, we know that a piece of good soft steel breaks with a fine crystalline fracture, and the same piece hardened when broken shows either an amorphous structure or one very finely crystalline, which would indicate that the crystals had been broken up by the action of heat, and that they had not had sufficient time to return to their original position on account of the sudden cooling. The tendency of the molecules of steel after hardening to assume their natural position when cold seems to be very great, for we have often seen large pieces of steel burst asunder after hardening, though lying untouched, and sometimes with such force as to hurl the fragments to some distance. If a piece of steel be subjected to a bright yellow or white heat its nature is entirely changed, and the workman says it is burnt. Though this is not actually a fact, it does well enough to express that condition of the metal. Steel cannot be burnt unless some portion of it has been oxidized. The carbon would of course be attacked first, its affinity for oxygen being greatest; but we find nothing wanting in a piece of burnt steel. It can, by careful heating, hammering and hardening, be returned to its former excellence. Then what change has taken place? I should say that two modifications have been made, one physical, the other chemical. The change chemically is that of a chemical compound to a mixture of carbon and iron, so that in a chemical sense it resembles cast iron. The change physically is that of crystallization, being due partly to chemical change and partly to the effect of heat. I have procured a specimen of steel showing beautifully the effect of overheating. The specimen is labeled No. 1, and is a piece of Park Brothers' steel (one of the best brands made in America). It has been heated at one end to proper heat for hardening, and at the other is what is technically called "burnt." It has been broken at intervals of about 1½ inches, showing the transition from amorphous or proper hardening to highly crystalline or "burnt." Malleable or wrought iron is or should be pure iron. Of course in practice it is seldom such, but generally nearly so, being usually 98, 99, or even more per cent. It is exceedingly prone to crystallization, the purer varieties being as much subject to it as others, except those contaminated with phosphorus, which affects it similarly with steel, and makes it very weak to cross and tensile strains. I have never estimated the quantity present in any except one specimen, a bar of 1½ round, which literally fell to pieces when dropped across a block of iron. It had 1.32 per cent. of phosphorus and was very crystalline, though the crystals were not very large. Iron which has been, when first made, quite fibrous, when subjected to a series of shocks for a greater or less period, according to their