The practical difficulties ought not, however, to be insuperable should gas-firing be otherwise found most practicable for the particular conditions at the smelter, although there appear to be certain physical characteristics of such flames which may be responsible for some of the difficulties met with in employing this type of fuel for the working of very large reverberatory furnaces. =The Condition of the Charge for Good Reverberatory Work.=—The considerations which decide the advisability or otherwise of installing at a smelter, any particular types of furnace, whether reverberatory or blast furnace or both, cover a very wide field, and will be more apparent when blast-furnace practice has been reviewed in detail. It is clear that the blast furnace is unsuited for the direct smelting of fine materials as such, and that the reverberatory form of furnace is best fitted for their treatment when large quantities of this material require to be dealt with. Actual practice has shown, however, that the reverberatory does not give equally satisfactory results on all classes of fines, and that there are certain physical and chemical conditions of the charge which appear to be necessary for the most successful and rapid smelting. When such conditions are not adhered to, less satisfactory working has resulted. Recent experience has, to some extent, defined more clearly the nature of these requirements, and has indicated the procedure which is necessary in order to avoid an undue supply of the less suitable material for the reverberatory charge. It is usual to smelt in the reverberatory furnaces, where such are available, the greater portion of the dust which accumulates in very large quantities in the flues at the smelter. The reverberatory is the only type of furnace in which such material could be treated directly, under the present conditions of working. In practice, however, it has been found in several instances, though not universally, that such dust is considerably more difficult to treat in the furnace, and entails considerably more expense in smelting than does the ordinary roasted concentrate. It is estimated by Ricketts that this extra cost is practically equivalent to the expense of roasting an equal weight of concentrate.