1 to 2 or 3 per cent. of phosphorus, beside some other impurities. An average of this cinder, both here and in Europe, would probably prove it to contain not less than 52 per cent. metallic iron. This class of cinder, in this country, is quite high in phosphorus, because the pig metal is boiled purposely very hot in order to throw the phosphorus as much as possible into the cinder for the better purification of the iron. Heating furnace cinder is not generally as rich as puddle cinder, still it averages a large percentage of iron, probably not less than 45 per cent., while some of it, taken from furnace bottoms, is even richer than puddle cinder. Forge cinder, the refuse of charcoal scrap sinking fires, and from Catalan forges, is a richer iron oxide than puddle cinder, and besides contains a very considerable quantity of reduced iron blended with it, which, by the usual method of working charcoal forges, it is impossible to separate. Roll scale is a pure oxide of iron, containing, as gathered up in the mills, about 70 per cent. of metallic iron. Although apparently little scale is made from iron in hammering and rolling it, yet in large mills the annual quantity, when summed up, would be quite surprising. Where puddling is carried on, scale is often used by the puddlers at certain stages of the boil, but even in many of these mills the scale is swept up and thrown with the cinder. Hammer and roll slag are both rich in iron, averaging perhaps nearly as much metallic iron as heating furnace cinder, say 40 to 45 per cent. It is quite worthy of note that all these refuse slags, so very rich in metallic iron, and produced in so very large quantities, are often wholly wasted, or, if utilized, only command a fraction of the market. value which the same number of units of iron would realize if con tained in ore. Most blast furnace owners would rather pay several dollars per ton for ores containing as low as 30 per cent. of metallic iron than to charge their furnaces with cinder, to be had for the hauling, that analyzed 54 per cent. of metallic iron. Owing to the alloy with phosphorus, and the very refractory nature of these cinders, furnace managers will at best only use them sparingly. Many who are proprietors of both rolling mills and furnaces prefer to waste their rolling mill cinder on the dump rather than use it in their blast furnaces, for fear of contaminating a superior grade of pig iron they aim to produce. Even in furnaces not very particular as to quality of iron, the use of cinder is usually limited to from 10 to 20 per cent. of the charge, the latter being rather an unusual quantity for ordinary brands of pig. Not less than 40 to 50 tons of puddle cinder daily made by the Philadelphia rolling mills is given away for the hauling and barely pays for the labor of transportation to cars a short distance from the mills, to be delivered to neighboring blast furnaces. By this process this wasted cinder could be daily converted into 12 to 15 tons of good wrought iron at low cost. In former years cinder pig iron was a marketable commodity, very much in demand for working into cheap rails, etc., but now ingot iron and low steel is so rapidly pushing it away, and displacing puddled iron in rolling mills, substituting superior for inferior stock, that cinder pig iron is no longer in request. With an increasing demand for better qualities, the less cinder can be used in blast furnaces, and the more it will accumulate. In Staffordshire and many other parts of England, at the present time, cinder has fallen so much in demand for furnaces that many rolling mills are seriously inconvenienced by the large accumulation of cinder around them. Indeed in some parts of Europe, and even in this country, cinder has been used to fill up vacant lots and low places, and sometimes even forms the embankment of railroads. In Sweden rich forge cinders, unused, have been accumulating in enormous piles for 50 years. I have now succeeded in working these cinders, so rich in iron, and rolling them to bars at one heat. I prepare this material with carbon and substantially the same cheap fluxes as used for ores. It is safe to estimate, under all contingencies, that one ton of muck bar can be made on a regular working scale from three tons of puddle cinder, deoxidized with bituminous or anthracite carbon in heats of from 500 to 1000 lbs., according to the size of the furnace. During a series of experiments at Round Oaks, Staffordshire, England, in the past winter, I produced as high as 43 per cent. by weight of muck bars in 24 hours from cinder and scale, deoxidized with bituminous carbon. In this instance the muck bar, with one-half scrap, has lately been cut up, sunk in charcoal, forged, reheated, and rolled to 1 inch wire rods, showing no greater waste than usually attends that mode of working iron. These wire rods were then reheated and rolled down to No. 6 wire. The wire was salt coated and in four passes was drawn to No. 12, after which it was annealed and meal coated, and in two more passes it was drawn to No. 16 wire. No better illustration than this can be presented to show the character of iron made from cinder by this method. The cinder from which this iron was made contained more than 2 per cent. of phosphorus, whereas an analysis of the iron showed it had been reduced to per cent. 100 Within three months, at the Phoenix Iron Company's works, 27,426 Ibs. of puddle cinder, 7350 lbs. of old bed Champlain Ore, and 2400 lbs. of iron scale was mingled with bituminous coal and slagging material. The combinations were varied, and it was then moulded into shapes for testing the system. The charges were placed in an ordinary double puddling furnace. Some of them were of cinder alone, some cinder and ore, and others cinder and scale. In this way a variety of tests were made, each with a sufficient quantity to run the furnace several heats upon one mixture. Part of the time the furnace was run day and night continuously, making then 6 heats in 24 hours, averaging 507 lbs. of muck bar at a heat. Since then the heats have been much shortened. The mixtures were all moulded 15 inches high into the annular cylindrical shape, being 83 inches outside diameter, and cored from top to bottom 3 inches diameter, and for more uniform heat circulation they were cross cored through and through at the base. With cinder mixture they contained each 48 lbs. of cinder, and of ore alone 52 lbs., while the weight varied between these points in proportion to the variation of the ore and cinder mixed together. In every one of these tests the iron oxide was reduced to metal, balled, squeezed in a rotary squeezer and rolled to muck bar at the same heat, presenting in lengths and appearance the ordinary muck bar from pig iron. The average period of the heats was 3 hours, while the yield of the whole in weighed muck bar was about 32 percent. of the weight of the oxide. At the previous test, before referred to, the heat was 24 hours and the yield larger. This is accounted for mainly by the lesser thickness of the moulded shapes, which admitted more rapid heat penetration.. The cost of moulding pipes of the less thickness by machinery is comparatively trifling in comparison with the quickened production and reduced cost of the metal. A greater advantage will be gained in doubling the hearth capacity to twice that of the ordinary double puddling furnace. Then 1000› lbs. of iron, with no more labor, may be as easily produced at a heat as 500 lbs. has already been done with the present size. Such increased furnace proportions would not be as economical for puddling pig iron, but as by this system the carbonic oxide generated by the metal mixture itself is largely depended upon as a heating power, and it is distributed over the entire hearth surface, the conditions are quite different from puddling pig iron. In this case the pure oxygen of the metal mixture from all parts of the hearth readily finds for itself (without furnace labor upon it) atoms of carbon in close contact with which to assimilate and generate heat. In puddling pig iron the 2 or 3 per cent. of carbon it contains is prevented from escaping by the melted metal. Its exit is sealed, or at least very much retarded, by the weight of the melted metal, and it is only after excessively laborious "rabbling" or stirring that the way is opened for its escape. Thus allowed to volatilize, it combines with oxygen into carbonic oxide, but as three-fourths of the furnace gas is nitrogen, for which carbon has no affinity, it becomes necessary to pass large volumes of air very rapidly through the furnace, in order to have sufficient oxygen present with which the carbon may combine. With the swift movement of these air currents, sweeping like a whirlwind through the furnace, much of the carbon from the pig is passed off undecomposed, and along with it vast volumes from the fuel, rendering the top of the stack a constant monitor of wasteful combustion. The following analyses have been made by the direction of David Reeves, Esq., President of the Phoenix Iron Company, by whose kind permission I use them. An analysis of the 27,425 lbs. of puddle cinder referred to as recently worked proved its composition to be The ore used was magnetic known as "old bed Lake Champlain," such as is usually used for fix, and highly charged with phosphorus. Fluxes and bituminous coal were mingled in different proportions. with a varied combination of cinder, scale and ore, and consecutively numbered. No. 1 was composed of 6000 lbs. of cinder alone. An analysis of the 1690 lbs. of muck bar produced from it showed it to contain No. 2 was a mixture of two-thirds cinder and one-third "old bed Champlain" ore, making 6000 lbs, in all. The 1644 lbs. of muck bar it produced analyzed Silica, . 45 +37 No. 3 was of 6000 lbs. of cinder alone, but fluxed differently from No. 1, and the analysis of its 1853 lbs. of muck bar was Silica,. 35 .38 No. 4 was composed of 1200 lbs. "old bed Champlain" ore and 4800 lbs. of cinder. The 1842 lbs. of muck bar from it analyzed No. 5 was a mixture of 2400 lbs. iron scale and 4800 lbs. cinder. The 2183 lbs. of puddle bar it produced analyzed Silica, . •29 .36 No. 6 was exclusively of 4673 lbs. of "old bed" ore, whereas the analysis of 1437 lbs. of puddle bar from it showed 62 •16 Silica,. No. 7, with a different mixture of fluxes from any of the others, was composed entirely of 1887 lbs. of cinder. The 537 lbs. of muck bar analyzed Silica,. 26 38 Some important conclusions may be deduced from these experiments. For instance, in the mixture No. 3, the 6000 lbs. of puddle cinder contained, by analysis, 177 of silica, or 1062 lbs. of the entire weight, whereas the analysis of its muck bar showed it contained only 3 of one per cent., or 63 lbs. This is an elimination of 1056 lbs. of silica from the 1853 lbs. of muck bar produced from this mixture. · 35 Again, the analysis of this No. 3 puddle cinder showed it contained 1.96 per cent. or 117 lbs. of phosphorus, while the analysis of the muck bar contained only 3 per cent. or 7 lbs. Here is an elimination of 110 lbs. of phosphorus out of 117 10 • 38 10 lbs. which the 1853 |