Now in a gas producer, by maintaining a sufficiently thick bed of glowing coal and admitting only such amounts of air as will produce mainly carbon monoxide gas, a product of high burning value is obtained. A kilogram (2.2 pounds) of carbon in burning from C to CO generates only 2450 calories or heat units, whereas its complete burning to CO_{2} would give 8080 calories. So by conducting the carbon monoxide gas—the product of the first stage of the combustion—through brick-lined pipes to the furnace, and in the latter by addition of air allowing it to burn to CO_{2}, the greater amount of heat (i. e., 8080 minus 2450 or 5630 calories) is evolved in the furnace. Of course, some of this theoretical two-thirds which is in this way made available at the furnace is lost because a little CO_{2} is formed, and always the nitrogen of the air used greatly dilutes the gas. But there are gains, notably the great heat which is carried over by the hot gas from the glowing bed of coal and that from the water-gas which is formed from steam used in the producer. So, all in all, the gas generated in a “battery” of gas producers, all of which discharge into one large main or header to maintain gas of average composition, is quite a satisfactory fuel. CHAPTER X CAST IRON From the preceding chapters we now know pretty well the place which cast iron occupies in the iron family. In the chapters which have succeeded the one in which we discussed the blast furnace and pig iron, every one of the products except crucible steel has been produced through some “refining” operation which greatly changed the composition, structure and properties of the product. Cast iron is not the result of a refining operation in this sense of the word. It is produced through simple mixing of pig irons of various compositions, usually with some admixture of iron castings of similar composition which have outlived their usefulness in the industrial world and have been returned as scrap to be remelted. When we say that cast iron is not produced through a refining operation, it must not be inferred that no change in composition occurs during the remelting. There is some change, notably a loss through oxidation from the air blast of a little of the silicon and manganese. Aside from this there usually will be absorption of enough, or sometimes more than enough, carbon from the coke used in melting to make up for the carbon which is oxidized. Usually some sulphur also is taken up from the fuel. There is, however, no such actual or intended alteration of composition through burning out of the metalloids as is necessary for the production of wrought iron and steel.