From the above it is seen that all sorts of cast iron can be produced ranging from the extremely hard, high cementite, white irons with low silicon content down to the very soft gray irons which result mainly because of higher silicon content and slower cooling. The white irons are more or less unstable as is shown by the decomposition through which the hard, white iron castings become “malleable” by annealing as was told in Chapter XII which discussed Malleable Cast Iron. [Illustration: No. 31. GRAY CAST IRON WITH FERRITE, PEARLITE AND GRAPHITE FLAKES (_Magnification 70 Diameters_) ] The gray cast irons are much more stable. They consist of what, in an early chapter, we referred to as “steels with an impurity, the graphite flakes.” They consist, then, of free, soft iron or ferrite, certain amounts of the characteristic steel constituent, pearlite, and the soft graphite flakes. They arrive at this composition, through breaking down of the austenite and cementite structures during the cooling,—just how not having been satisfactorily determined. Consistent study is being put upon this subject and several unique and long-studied possible explanations of this section of the full equilibrium diagram have been proposed and debated, all based upon the data so far available. While much information on this subject has been gained the matter is still so much in dispute that it is best for us to venture nothing definite in regard to just how the changes occur. The reference books for this chapter (see page 354) give quite completely the data, theories and explanations so far available. Compared with the steels, the cast irons are vastly complicated. In them we have elements which occur in practically negligible amounts in the steels. Commercial cast irons, for instance, have silicon ranging anywhere from ½% to 3%, phosphorus .10% to 2%, graphite 0% to 3.50%, and carbon in the combined form (pearlite or cementite) from 3.50% to .10%. If these represent the majority of cast irons what about our pig irons which have 2, 5, 8, 10 or even 15% of silicon, others with 1 or 2 and occasionally very much more of manganese, and still others which vary widely in phosphorus content?—for from the metallographic and physical chemistry standpoint pig irons are cast irons. As we can never get perfectly pure iron-carbon alloys to experiment with, their content of other elements, silicon, nickel, phosphorus, etc., vitiate more or less the results obtained, but even could such pure alloys be secured we are not greatly helped since our serviceable alloys are never such. Each added element brings about greater complication and one does not wonder that in the short twenty years which have elapsed since study was seriously undertaken, metallurgical science has not entirely solved the big problem. REFERENCES GENERAL & VARIOUS “The A B C of Iron and Steel,” Penton Publishing Co., Cleveland.