hard casting which will allow it to anneal properly or not. First the chemical composition of the metal itself. Second the thickness of the sections of the casting, and third the pouring temperature. The last two items have been gone over above. It remains to give specifications for the first. The most powerful agent affecting the state of the carbor present in a casting is the silicon. As it is necessary to have a casting white in fracture as it leaves the sand, the silicon must be very low. Then with the proper pouring temperature, and when poured into sections suitable for the composition employed, the results will be good. Naturally this will principally depend upon the thickness of the work made, as this cannot be changed, and then upon the heat of the iron, which can be regulated by careful melting. The thinner the castings, the higher the silicon the mixture can stand. Thus for pipe fittings, the silicon may run up to 1.00 per cent. in the casting. For couplers, and other heavy railroad castings the silicon has to run down as low as .45 per cent in order to get the best results. When charcoal irons were used exclusively (these standing more punishment in melting than the coke irons of the present day), the silicon oftentimes ran as low as 0.28 per cent. in a casting and still this was first-rate. The general average, how ever, for all around medium and fairly heavy work is 0.65 per cent. silicon in the casting, which means about 0.85 per cent. to 0.90 per cent. in the mi ture. 0.45 per cent. may be considered the lowest range for heavy work, and 1.00 per cent. the highest, for the lightest of castings.

The phosphorus should not exceed 0.225 per cent., the manganese not over 0.30, the sulphur as low as possible, preferably not over 0.05 in the casting, though in Europe, where the long anneal counteracts this evil, the sulphur goes very high, sometimes even up to 0.30 per

cent.

The lower the total carbon, down to 2.75 below which trouble arises, the stronger will be the casting. Hence steel scrap is added to make the metal low in its carbon content. This is a much better plan th. to refine the iron in the process to get the carbon low. In general it is best to simply melt a mixture, and then get it out of the furnace as quickly as possible, in order to get it away from oxidizing influences as quickly as may be. Five to 10 per cent. of steel may be added, also malleable scrap, if necessary. In a 10-ton heat the best proportion of the mixture is five tons of pig iron, one ton malleable scrap, 500 pounds steel scrap, and the balance the sprues of the previous work. The practical effect of these steel and other additions are about as follows: 100 pounds wrought iron scrap equal 250 pounds steel, equal 2,000 pounds malleable scrap. Mixtures thus arranged come out about the same in strength, all other things being equal.

Charcoal iron is now being used only where the source of supply is close, and its cost is but 50c. to a dollar over the best "Coke Malleable" or "Bessemer Malleable." as coke irons made specially for the malleable foundry are called (the last named being a Bessemer iron with the phosphorus a little higher than is allowed for steel). It is best to use these

classes of "malleable" as pig irons, for they are made with an extra amount of coke, and are much less oxidized than the irons blown under high pressure and very hot blast.

The mixtures used in the malleable foundry are as follows: Where the cupola is used for malleable castings in addition to making pots for annealing, the regular mixture as used for the air furnace or open hearth furnace must be reduced slightly in silicon, as the cupola burns out less of this element. Hence about 0.25 per cent. is to be added to the amount of silicon required in the casting. In the case of the furnaces, about 0.30 per cent. to 0.35 per cent. must be added. For pots it is advisable to use good pig irons, and to utilize salamanders. and the large scrap pieces that are unsafe to put into the furnace. The silicon in the pots as cast should be about 0.60 per cent., and hence about 0.85 per cent. should go into the cupola.

The mixtures for air furnace and the open hearth furnace are about the same. Possible the air furnace should have a little more silicon, as it is under the influence of the gases from co bustion longer. There is also necessary an occasional use of ferrosilicon, especially in the open hearth, as through accident, the metal may be badly burnt and the addition of ferrosilicon brings about the proper composition, though the metal should not be put into castings, but cast into pigs, to be fed subsequently into the regular mixtures in small quantities at a time. About 250 to 500 pounds ferrosilicon is usually all that is necessary for this purpose. actual amount can be calculated at the time from the supposed loss of silicon in th bath.

The

Malleable castings are made in the cupola, the air furnace, the open hearth furnace, and in the crucible. The last named process is now only practiced in Europe, being too expensive in this country, though turning out a most excellent product. The cupola process makes the poorest castings, as the metal is in contact with the fuel in this method. Hence the absorption of sulphur, and oxidizing influences which are partly avoided in other processes. The peculiar constitution of the metal as cast makes it necessary to anneal it at a temperature some 200 to 300 degrees F. higher than ordinary air-furnace iron, which means a greater expense for wear and tear on the ovens and annealing pots. This class of castings is therefore only used for the cheaper grades of malleable castings, such as pipe fittings, and hardware castings, where great strength is not essential, and enough ductility is had to satisfy the demands made on the work. The selling price is also about half a cent a pound less than the high grade metal.

The bulk of the malleable castings made in this country come from the air furnace. This is an excellent melting process, can be manipulated easily, is not too expensive, and will probably continue to be the method used in most of our malleable works. The air furnace, as used for malleable purposes is somewhat different than that used for making rolls and gun castings. It is illustrated herewith.

The entire roof can be taken off in sections, called "bungs." so that the sand bottom can be made, and the charge put in. Where more.

FIG. 2-Section Through Fire-Box of Air Furnace for Malleable Castings.

quickly, as the metal is continuously oxidizing while in the furnace, after it has reached the proper composition. A ten-ton heat, when poured into small castings often takes three quarters of an hour to tap, and hence the first iron and the very last may be two different things. Hence the dividing up of the work to get the class of metal best suited to the castings to be made. The amount of coal used to melt iron in a well constructed air furnace is four pounds iron to one pound coal. In the case of the cupola, while ordinary gray iron practice requires one pound coal to every eight pounds iron, in malleable work, it takes one pound coke to only four pounds iron, or just the same as good air furnace practise. It is but just to say, however, that in many foundries of the country, the air furnaces are so poorly constructed that oftentimes one pound coal melts only two pounds iron.

Bottom is made placing layers of fire sand, that is sand with about 98 per cent. silica, and very free from fluxing impurities, on the brick

quality of the bituminous coal used, and the construction of the furnace. From six to eight heats can be made on the same bottom, with but little repairing, but the usual run is from two to four. In order to know when the heat is ready, a test plug, so called, is cast. This plug is about of a diameter equal to the heaviest section of the castings to be made. It is about eight inches long, and the mold for it is made by simply pushing the pattern into the sand in a box. The metal is taken from the bath by a small ladle dipped into it as deeply as possible. After pouring, as soon as the iron is set, the plug is grasped with a tongs, is dipped into water to cool it, and then broken across. The fracture is observed, and if properly crystalline, and with but little or no mottling, the heat is ready to tap. If there is too much mottling, that is too much graphite left, the process is continued to burn out more silicon, and also get the metal hotter, and another test plug taken. Experience will tell just how long a heat still has to run until ready to tap. After the iron has melted, the slag is skimmed off, and this gives a good chance for refining_action, with means the burning out of silicon. The test plug is always taken after skimming, which is often done for the second time. When the heat is tapped, the men take it off in hand ladles, and pour the molds, throwing the iron into them as quickly as possible so that the necessarily small gates do not prevent the metal from filling the molds by chilling and resulting in "short pours." The open hearth process (See STEEL CRUCIBLE PROCESS) is by far the best one in general use, but is confined to those works where great quantities are made year in and year out. Thus there are several works where about 80 tons of castings are made daily, and in which the use of the open hearth is a paying proposition, especially as the same furnace can be used for making acid steel heats in place of malleable cast iron, as desired. The fuel consumption for the open hearth corresponds to one pound coal for six pounds iron melted; showing a considerable economy over all other methods. The use of a gas producer system, however, where

natural gas is not available, makes the installation an elaborate one, and not desirable where the proper class of help is not available.

In the case of the open hearth, the furnace is always hot, and hence a heat is finished about an hour sooner than in the air furnace. The iron gets hotter, and can be taken cut in five ton ladles to be distributed afterwards and as the metal is not as long in contact with the gases as in the air furnace process, it is of better quality. The most economical size of furnace is the 20 ton, with the crane ladle to take off the metal in large quantities, so that tapping is not so long continued a matter as in the air furnace. The latest patented invention to assist in this is the application of two or more spouts to the furnace, so that metal may be taken out at different levels. In this way the surface of the bath, which is punished most by the gases, may be taken off first. Then while this is being poured off, the next part of the heat, now the top, is again taken off, and finally if three spouts are used, the bottom may be

the oven. In this way they will not be cracked before going into the annealing ovens. The modern tendency is to introduce the sand blast for cleaning, as this removes every particle of sand from the hard castings. Where the iron has been allowed to get too low in silicon, or "high" as it is called, in contradistinction to "low" iron, where the silicon is too high, and the metal mottled or even gray; the sand is apt to burn on so hard that the tumbling all day does not remove it all. Here the sand blast is excellent.

From the sorting room the castings go to the annealing room, or rather to a part of it in which the packing is done. To anneal the hard castings they are placed into so-called "saggers" or annealing pots. These are simple box like shells, with no bottom, about one inch thick, and say 18"x24"x15" high. Three or four of these are placed over each other, and on a "stool" high enough to allow a free circulation of the gases under it. The castings are carefully placed in these pots and packed with "scale" in

taken out as long as half an hour afterwards without any deterioration or change in the metal.

From the foundry the castings, after shaking out the molds, go to the hard tumbling room, where they are freed from the adhering sand. This is done by means of tumbling barrels into which the castings, and a supply of "stars" made of the same hard iron, are placed. Where castings are liable to crack by this tumbling about, sticks of wood are introduced, so that as they strike them, no damage is done. Where delicate castings are made, these are pickled in dilute sulphuric or in hydrofluoric acid.

After cleaning the hard castings, they are carefully sorted out, the cracked or otherwise defective pieces thrown out, to go back into the furnace again, and the good ones are sent into the annealing room. Where castings are made which crack as soon as they cool, on account of their shape, such as the hand wheels of freight cars for braking purposes, these while still red hot in the sand, are taken into small ovens where they are kept quite hot for a time, and are then allowed to cool very slowly with

such a way that when red hot, the whole may not settle and warp the work. This scale is puddle scale, hammer scale, or even iron ore. For that matter, as the process is more of a conversion of the combined carbon into the "temper carbon" the castings can be packed in fire clay or sand and good results obtained. But the puddle scale seems to give the best results, with greatest cheapness. The flakes that fall from the annealing pots, these lasting only for 7 to 14 heats, can be crushed and make the finest kind of packing material, being pure oxide of iron, and no further scale than the initial lot need be purchased.

The pots, properly filled, are covered with a "mud" made of the sand rolled off the hard castings mixed with water. The joints of the pots are also carefully daubed up with this mud; the pots are introduced into the oven either run in by a special carriage in the old style ovens, or lowered in from the crane in the new ovens, the tops of which can be removed. The ovens are now fired and within 36 to 48 hours, the full temperature of 1,350 degrees F. in the coldest

After the pots are taken out, they are dumped, the castings taken from the scale, and the former sent to the soft tumbling room. Here they are packed with small pieces of annealed mallable castings, or with wood and leather, such as old shoes, etc., so that the adhering scale is removed, and a fine coat of graphite is given them. They come out shining black, and can be shipped direct, or else coated with asphalt, as required. Sometimes the castings must be straightened, which should always be done cold, either by drop hammer or hydraulic press. Test plugs are usually cast on the important work, so that breaking these off on inspection, the quality of the metal in the particular casting is revealed at a glance. Test bars for physical test should also be taken off with the first and last portion of every heat, so

000 tons annually. In Europe about 50,000 tons are made in the same time. Most of the large companies now make steel castings in addition, as malleable cast iron though most excellent for shock, is not strong enough to stand under the terrific strains that are now put on the structures where this work is used. Thus the 100,000 lb. cars now require steel couplers, as the malleable ones tear apart, yet the latter will stand all the bumping that comes along, while steel will not. The principal use for malleable castings is for railroad work. Next comes agricultural machinery. After that a great variety of work, such as pipe fittings, hardware, machine parts. pistols. tools, etc. The demand is constantly growing, and while steel was supposed to be driving it out, this has not turned out to be so, but for every casting that must be made,