_Feeding the Charge._—It is self-evident that, the more thorough the preliminary preparation of the charge before it reaches the zone of fusion, the more rapidly can the actual smelting proceed. A piece of raw ore that finds itself prematurely at the tuyeres, without having been subjected to the usual preparatory processes of drying, heating, reduction, etc., must remain there until it is gradually dissolved or carried away mechanically in the slag. Any such occurrence must greatly retard the process. It would seem, by the same reasoning, that an intimate mixture of the ingredients of the charge should expedite the smelting, and I advocate the intimate mixture of the charge ingredients in all cases. The theory of feeding is simple, but not so the practice. If the charge column were composed of pieces of uniform size, the ascending gases would find the channel of least resistance close to the furnace walls and would take it preferably to the center of the shaft. The more restricted channel would necessitate a higher velocity, so that not only would the center of the charge be deprived of the action of the gases, but also the portion traversed would be overheated; many particles of ore would be sintered to the walls or carried off as flue dust; slag would form prematurely; fuel would be wasted; in short, all the irregularities and losses which accompany over-fire would be experienced. In practice the charge is never uniform, but is a mixture of coarse and fine. By lodging the finer material close to the walls and placing the coarser in the center, an adjustment may be made which will cause the gases to ascend uniformly through the smelting column. A furnace top smoking quietly and uniformly over its whole area is the visible sign of a properly fed furnace. _Effect of Large Charges._—It has frequently been remarked that, within certain limits, large charges give more favorable results than small ones; and numerous attempts have been made to account for this fact. My observations lead me to offer the following as a rational explanation—at least in cases where ore and fuel are charged in alternate layers. Large ore-charges mean correspondingly large fuel-charges. The gases can pass readily through the coke; and hence each fuel-zone tends to equalize the gas currents by giving them another opportunity to distribute themselves over the whole furnace area, while each layer of ore subsequently encountered will blanket the gases, and compel them to force a passage under pressure, which is the manner most favorable to effective chemical action.