The coke used should be of good quality; the formation of a fused ash (clinker), in any quantity, causes ceaseless trouble, and requires frequent removal. The coke should be broken into lumps of a uniform size (about 2 in. across) before being brought into the office. The furnace should be well packed by stirring, raising the coke and not ramming it, and it should be uniformly heated, not hot below and cold above. In lighting a furnace, a start is made with wood and charcoal, this readily ignites and sets fire to the coke, which of itself does not kindle easily. In commencing work, add (if necessary) fresh coke, and mix well; make hollows, and into these put old crucibles; pack around with coke, so that the surface shall be concave, sloping upwards from the mouths of the crucibles to the sides of the furnace; close the furnace, and, when uniformly heated, substitute for the empty crucibles those which contain the assays. It is rarely advisable to have a very hot fire at first, because with a gradual heat the gases and steam quietly escape through the unfused mass, while with too strong a heat these might make some of the matter in the crucible overflow. Moreover, if the heat should be too strong at first, the flux might melt and run to the bottom of the crucible, leaving the quartz, &c., as a pasty mass above; with a gentler heat combination is completed, and the subsequent fiercer heat simply melts the fusible compound into homogeneous slag. The fused material may be left in the crucible and separated from it by breaking when cold. It is generally more convenient to pour it into cast-iron moulds. These moulds should be dry and smooth. They act best when warmed and oiled or black-leaded. Air entering through the fire-bars of a furnace and coming in contact with hot coke combines with it, forming a very hot mixture of carbonic acid and nitrogen; this ascending, comes in contact with more coke, and the carbonic acid is reduced to carbonic oxide; at the top of the furnace, or in the flue, the carbonic oxide meeting fresh air, combines with the oxygen therein and re-forms carbonic acid. In the first and third of these reactions, much heat is evolved; in the second, the furnace is cooled a little. It must always be remembered, that the carbonic oxide of the furnace gases is a reducing agent. When these gases are likely to exert a prejudicial effect, and a strongly oxidising atmosphere is required, the work is best done in a _muffle_. FOOTNOTES: [3] There is an exception to this, as when the slag is liable to be acted on when exposed to the air and to the gases of the furnace. In this case a layer of fused common salt floating on the slag, so as to protect it from the air and furnace gases, is a distinct advantage. CHAPTER III. WET GRAVIMETRIC METHODS.