This property is made use of when it is desirable to obtain hard wearing surfaces, and in the same casting tough and soft central portions, as in car wheels.

While this chilling effect is exceedingly valuable for many purposes,

not allow the coldness of the mould to affect the grain.

(5) The thickness of the sand scale.

This is an important consideration; for the sand must be cleaned from the casting, and the surface must first be cut before the interior can be reached.

To prevent the iron from burning the sand into itself and thus forming a scale, a plumbago facing is sifted on the surface of the mould, but it is difficult for the facing to lie on the surfaces or to resist the intense heat of the metal.

When aluminum in an iron causes the dropping of the graphite from the mass of the metal, that graphite which is on the surface of the casting separates and forms a perfect plumbago facing, which opposes the sand and the heat. It will therefore be seen that in castings having sufficient aluminum to cause this separation of graphite, there will be no sand clinging to the face, and that the surface will be as soft as the interior of the casting. Every iron worker will appreciate this good effect of aluminum.

(6) The effect upon hardness.

Hardness in cast iron is caused by the carburetted or white iron, in masses large enough to oppose the tool. If the carburetted iron exists in minute threads stretched around atoms of graphite, a tool will easily cut it and it will not be considered hard.

This graphitic carbon, minutely dividing the mass, gives the tools of the workman a chance to cut or break the films of metal, giving what we call softness to the iron. The later the carbon is dropped, the smaller will be the atoms of graphite and the closer the grain. Yet this greater subdivision will, for the reason just given, make the iron work more easily.

The fineness of the grain of iron affected by aluminum causes such iron to be much more easily cut than iron of coarser grain.

The next question to consider is that of strength.

The power of wrought iron and steel to resist extension is so great that where such stresses are to be resisted, decarbonized metal should be used. The resistance of any cast iron to crushing is so great that we need not consider this.

The forces which cast iron structures should be made to resist, aside from crushing, are a dead weight or a blow applied transversely. We should therefore test cast iron with these forces.

(7) The resistance to a load gradually applied, or a dead weight.

CHART 3.

If we compare the transverse breaking weights of the two series which we have been considering, number by number, we perceive that the aluminum has increased the strength to sustain a constant load.

This is a very important effect and perhaps comes partially from the tenacity and strength of aluminum itself, but probably more from the uniform grain of the iron.

CHART 3.

(8) The resistance to a load suddenly applied, or impact.

It may be thought that the effect is substantially the same, whether the force be a constant weight or a suddenly applied blow. We shall at a future time prove that the effects are not the same and that an iron should be tested by a blow if it is expected to resist impact.

CHART 4.

By a comparison of the graphic representation, we see that the capacity to resist impact is increased by the addition of aluminum much more than the capacity to resist a dead weight. It will be seen at a glance that the test bars made with the white base are benefited far more than those made with the gray base. The reason for this is, that the white base alone made porous castings; at each remelt this porosity increased, due to the continuation of the heat, running the strength down to 68 pounds at the fifth heat.

The first, and each subsequent addition of aluminum, caused the castings to be perfectly sound, and the infinitesimal atoms of graphite deposited throughout the metal removed the rigidity and brittleness of the initial metal.

The gray iron base contained enough silicon to accomplish all this, and the only effect on strength that the action of the aluminum on carbon could have, would be to increase the fineness of the grain unless the toughness of the aluminum itself could give strength to the casting, though the aluminum no doubt removed any slight blow holes that existed in the initial gray metal.

CHART 4.

This leads us to notice that each addition of aluminum increases the strength over that of the initial metal. We must expect, however, that after we have added enough aluminum to cause a solid casting, and to remove the brittleness, that the dividing up of the mass by the atoms of graphite accomplishes, any further additions of aluminum

and consequent increase of graphite (which has no strength of itself) must weaken the casting.

(9) The elasticity.
CHART 5.

The compactness and closeness of the grain of cast iron when aluminum was the agent by which the graphite was precipitated, and the fine attenuation of the veins of the carburetted iron, cause the metal to be very elastic and, as we have seen, not so brittle as without aluminum.

(10) Permanent set.

WHITE BASE

This is caused by the compression of the graphite within the framework of carburetted iron. When this compression of graphite carbon is produced by transverse bending, the framework of the metal also takes on a permanent form which cannot be altered except by a greater force than was before applied.

The fineness and compactness of iron alloyed with aluminum gives less permanent set than iron equally as

DEFLECTION.

.35

.30

.25

.20

.15

.10
.05

GRAY BASE

0441044 1 2 3 4

soft, when such softness is produced by silicon.

U

11

ALUMINUM

CHART 5.

(11) The effect on the shrinkage of the iron.

The more suddenly and completely the carbon is changed from combined to graphitic, at the instant of crystallization, the more space will the casting occupy. When the casting is cold, it will therefore have contracted

less than if more carbon had remained combined. White iron, having most of its carbon in the combined state, shrinks from one-fourth to onethird of an inch in each foot.

Gray iron sometimes shrinks as little as one-tenth of an inch to each linear foot. As the combined is the natural state for the carbon, we may say that this maximum shrinkage is the natural shrinkage for cast iron having its carbon combined. We can therefore say that aluminum takes out or reduces shrinkage when a sufficient quantity is added. This is a very great advantage, as shrinkage requires great skill in the preparation of patterns to prevent warping and cracking, and violent internal strains within the castings.

The lessening of shrinkage avoids these evils, and is therefore a great gain.

CHART 6.

If you will look at the chart for shrinkage, you will see the most conclusive proof of our explanation of the way in which shrinkage is lessened. With both the white and the gray bases, during the first two additions, the shrinkage of the square bar is slightly increased.

The influence of the aluminum thus far has been in the direction of elimination of blow holes, and causing an even distribution of the dark and light grains.

CHART 6.

At the third addition, however, when the amount reached three-fourths of one per cent, the effect was appreciably felt upon the carbon, as seen by the color, and as we should expect from the deposition of this large bulk of graphite, the casting does not shrink as much, and each addition of aluminum increasing this bulk of graphite decreases the shrinkage.

CHART 7.

The effect upon the grain and color of the thin bars of the series is very remarkable, showing that the aluminum has changed enough carbon to

the aluminum increases, but in the series for comparison the shrinkage dropped still more rapidly.

If a new crucible was used in commencing this comparison series, enough silicon might have been absorbed to produce this effect. This leads us to remark that on account of the variations of conditions in any series of tests, that cannot be foreseen, we must avoid drawing any but general conclusions, and these should be based upon a large number of experiments.

(12) The fluidity of the melted metal.

CHART 8.

Our tests of fluidity are correct as far as each individual heat is concerned, but variation may be due to the heat of the metal of that particular cast when poured. Viewed in a general way, the indications are that with the white base, with almost no silicon, the aluminum has increased the fluidity; but judging from the series with the gray base we would say that combined with silicon aluminum reduced the fluidity.

FLUIDITY.

Our remarks in connection with shrinkage show that a sharp casting is

will be too small to have much effect upon the fluidity of the metal. The fact of the iron giving sharper and more perfect castings, on account of the swell of the casting, caused by the deposition of graphite at the instant of solidification, might cause the iron to be pronounced more fluid, if judged by the appearance of the castings.

No doubt the presence of varying quantities of manganese, sulphur, phosphorus and silicon, in the cast iron used, would modify the influence of aluminum, and until this is understood it may require considerable experiment to determine the amount of aluminum required, or how it shall be introduced.

This hurried presentation of the remarkable effects of aluminum upon cast iron will give an idea of the great benefit which is now promised to the iron founder by the rapidly falling price of aluminum as cheapened by the electric furnace.

We have already occupied all the time that was set apart for us.

Following the publication of this part of the subject, we shall soon present the results of the laboratory work of Professor Mabery and Mr. Vorce, which will throw still more light upon this interesting subject.