The gas generator is tended in the usual way with the ordinary precautions. To keep ashes and dust out of the gas tube, lumps of charcoal are heaped up to the height of the top of the flue. The wind pressure for the generator was 33 to 41 millimeters of mercury, that of the wind for the combustion of the gas (at Finspong the blast is not heated) being only 16½ millimeters. The pressure of the gas in the tube near the pot was 6.2 millimeters of mercury. The method of working, viewed chemically, does not sensibly differ from puddling; although giving as good, perhaps better, results at a much less cost. There are three principal periods in the operation: 1. The period before boiling. 2. The boiling itself. 3. The end of the boiling, and the formation of balls. When cast metal is poured into the pot a shovelful or two of refinery slag is added. The temperature of the bath is thus brought down; it thickens and boils, the pot revolving at the rate of 30 or 40 revolutions a minute. The metal is worked with a rabble, either to cool it or to get the slag to incorporate with it, as is done in puddling. Note must be taken of the temperature of the melted metal and that of the pot, at the moment of charging, the heat during working being regulated accordingly by increasing or diminishing the inflow of air and gas. When circumstances are favorable, boiling begins five minutes after the metal is run into the pot, and it lasts about ten minutes. Boiling having begun, the batch swells, the iron forms, granulates, and seems to cling to the rabble and the sides of the pot. The rotation of the pot is continued, as well as the working, to separate out parts which are not yet refined; but no more cold cinder is put in. While boiling goes on the temperature is regulated so that the pig does not cling to the side of the pot during a complete revolution, but so that the particles next the side fall back into the bath when the side comes uppermost in the revolution. The heat is raised a little when the iron can be felt by the rabble to be completely refined, when shining lumps make their appearance in the bath, and the iron begins to cling to the walls. At the moment, therefore, that the temperature is brought to its highest point, and the iron begins to agglutinate, the rotation of the pot should be stopped, and either immediately, or after the delay of a couple of minutes, it is removed. If the iron does not ball well, it is not completely refined, and the pot may be started again. If the iron is firm enough already, the isolated particles are exposed to the hottest flame possible, the blast being carried to its maximum. The refining is thus completely finished, and all the particles are agglomerated. The mobility of the gas tube at _m''_ is of advantage in this operation. It is sometimes useful to start the pot again to round up the puddled ball, but it is best if this has been formed with the rabble.