The greater portion of the dust results from the treatment of unsuitably fine material in the blast furnace, and by decreasing the quantity of this constituent the flue-dust problem will be largely overcome. The smelting of fine concentrate in the blast furnace has up to the present been considered judicious where circumstances have rendered imperative the addition of sulphides to the charge irrespective of their physical condition (either to act as a base for the matte, or on account of their fuel values), though naturally the proportion of fines has been kept as low as possible. The recent developments in sintering processes, however, suggest the possibility of the future successful treatment, after preliminary agglomeration, of fine concentrate in the blast furnace, and if it be found possible to conduct the sintering by utilising the heat of oxidisation of the more free sulphur atom of the pyrites, and thus leave the bulk of the iron-sulphide fuel values in the sintered product, as suggested by Peters, the difficulties in connection with excessive flue-dust production from the above causes will be largely overcome, and the reverberatories will thus be relieved of this difficult constituent of their charge. It therefore appears desirable, when circumstances permit, either to agglomerate fine concentrates and then treat them in the blast furnace, or else to roast them and smelt the product in the reverberatories. So far as present experience has gone, it appears that—other circumstances being equally favourable—the correct scheme of treatment depends almost entirely upon the composition of the concentrate, there being for each process a particular class of fines for which it is best suited. The sintering process deals most satisfactorily with one class of concentrate, whilst the roasting process seems more particularly suited for a different type of material. Thus the higher the iron and sulphur values, and the lower the silica content, the more successful, cheap, and efficient is the roasting process—the Anaconda material for example roasts well, requires practically no external fuel or heating, and with the added flux, works very successfully in the reverberatories. As the silica content increases, however, and the iron and sulphur contents diminish, there is a consequent decrease in the natural fuel values of the material, and as a result, the roasting is neither so efficient nor so cheaply operated, owing to the need of external fuel for giving the required roasting temperatures. On the other hand, it appears to be just this class of material which is best suited for blast-roasting.