temperature that produces the greatest softening, and the properties of the tempered steel will depend upon the precise degree to which the metal has been reheated. For example, if the product be required for making into sword blades, or watch-springs, and to possess much elasticity, the proper temperature is between 550° F. and 570° F.; but if the steel is to be suitable for saws the temperature must range within a few degrees of 600° F., according to the fineness of the tool intended; a lower temperature would give a metal too hard for them to be sharpened with a file. On the other hand, sharp cutting instruments and tools for working metals are obtained hard by tempering at lower degrees than springs. In practice the index of the temperature is taken from the colour of the film of oxide that gradually forms on a polished surface of the metal as the heat is raised, and begins by a very pale yellow (at 430° F.), passing through deeper shades into brown, then through purple into deep blue (at 570° F.), etc. The reader will now see why watch and clock springs have their deep blue colour, and he can observe for himself the whole series of colours by very gradually heating a piece of polished steel over a small flame. If we compare the chemical composition of wrought iron and of cast iron with that of steel as regards the content of carbon, we see at once that steel holds an intermediate position, so that if in the puddling furnace we could arrest the decarbonization at a certain point we should obtain steel; or if, on the other hand, we could put back into chemical combination with the decarbonized wrought iron a due percentage of carbon we should in that way also obtain steel. And it will be observed that the oldest primitive furnaces could not have failed sometimes to have produced steel as the net or final result of such actions. In fact, steel always has been and still is produced on one or other of these two principles, applied in divers ways, but severally and distinctly directed to that end. Of the many more or less modified processes of steel-making that have been in use, we need here but briefly mention a few which were _the_ processes of the first sixty years of our century, and are to a considerable extent still in operation, although eclipsed in importance by two other processes that, since the date referred to, have been supplying the metal in enormously increased quantities, and which will have to be particularly described.