the slow burning powder dissipating themselves into space. The tamping and fuse were often thrown out with great force, as from a gun, while little effect would be produced on the lead cylinder. Our tests, then, merely show the rupturing or rending strength of the powder with dry tamping in a short length of tube, while no measure of the projectile force is obtained. In the higher-grade dynamites there was apparently very little projectile force; with the lower-grade dynamites there was more; and in the black powders it was very marked.

Table VII. exhibits the results of the tests made with the Hazard, Oriental, Laflin & Rand, American and DuPont powders; the samples of the first being sporting powders only, while of all the rest, blasting as well as sporting powder was tested. The final column gives the price per pound of each of these explosives.

Messrs. Williams and Seager worked industriously for about four months on the preceding tests. This manner of testing gives consistent results where the powders tried are homogeneous and well mixed, and if tried with larger lead cylinders, one-ounce charges or more, and a greater variety of powders, the results ought to be very valuable.

THE PROPERTIES OF ALUMINUM, WITH SOME INFORMATION RELATING TO THE METAL.

BY ALFRED E. HUNT, JOHN W. LANGLEY AND CHARLES M. HALL, PITTSBURGH, PA.

(Washington Meeting, February, 1890.)

THE PURITY OF COMMERCIAL ALUMINUM.

A GREAT deal that has been written heretofore about the properties of aluminum is of doubtful value, owing to the lack of knowledge we have of the purity of the aluminum referred to. Much of the metal heretofore sold in the markets as pure aluminum has been contaminated with from 4 to 10 per cent. of impurities. Indeed, it is only within a very few years that a purer metal than 96 per cent. aluminum has been made upon any larger than a laboratory scale.

The works of M. Pechinet, of the Société Anonyme d' Aluminium, at Salindres (Gard), near Marseilles, in France, have long enjoyed the

reputation of making the purest aluminum placed upon the market. The method employed is essentially Deville's, reducing aluminum chloride by aid of metallic sodium, as practiced also by the Aluminium Company, Limited, at Oldbury, in England, and (with the modification of using aluminum fluoride instead of aluminum chloride, to be reduced by the sodium) by the Alliance Aluminium Company, of Wallsend-on-Tyne, in England.

The Pittsburgh Reduction Company, manufacturing under the Hall process, reduces the metal from its oxide, alumina (Al2O3), by electrolysis, this alumina being held in solution by a molten fluoride bath, which itself is not decomposed by the electric current. The latter is conveyed to the melted solution by means of carbon cylinders placed in the bath for positive electrodes, a carbon-lined metal pot forming the negative electrode. The oxygen of the alumina goes off at the positive electrode as carbonic acid, wearing away the carbon at the rate of nearly a pound of carbon to the pound of aluminum produced. The reduced metal settles to the bottom of the pot, from which it is easily tapped or ladled off, practically free from the electrolyte. Remelting entirely purifies it.

All four of these concerns have succeeded, although only with careful selection of materials, and the greatest care to prevent contaminations, in making aluminum of over 99 per cent. purity, and in regularly placing upon the market in considerable quantities metal of over 98 per cent. purity. They are the only establishments known to us which have done so.*

The impurities of the metal made by the sodium process, as practiced by the first three, have been nearly half iron and half silicon. In the case of the Pittsburgh Reduction Company the impurity consists almost entirely of silicon, the iron being less than ten-hundredths of 1 per cent.

In the first three-quarters of the month of December, 1889 (the plant was shut down for repairs, the first time since November 28, 1888, from the day before Christmas until New Year's), 1250 pounds of aluminum was made by the Pittsburgh Reduction Company, all of which contained over 98.25 per cent. aluminum, with less than 0.10 per cent. iron, and with about 1.50 per cent. silicon. A con

* Since this paper was first printed, we have received a communication from the Swiss Company at Neuhausen, protesting against the assumption that its product is fairly represented by the analysis given in Table I. This analysis was made upon a sample obtained at Neuhausen by Mr. Hunt, who saw it cast. We do not, however, insist upon it as a fair average; and we shall be happy to publish the analysis of any further sample.

VOL. XVIII.-34

siderable quantity of this metal contained over 98.75 per cent. aluminum.

A singular fact is that no small portion of the silicon in aluminum exists in the graphitoidal state. This was pointed out by Mr. H. J. Williams, then chemist of the Pittsburgh Reduction Company, in Transactions, vol. xvii., page 542, and was recognized by Deville also in his researches. In aluminum containing 6.20 per cent. of silicon, made by the Alliance Aluminium Company, Limited, we have found 3.85 per cent. of graphitoidal silicon. In a sample of aluminum made by the Pittsburgh Reduction Company from calcined native bauxite, and containing 5.30 per cent. of silicon, 2.40 per cent. existed in the graphitoidal state. This graphitoidal silicon can readily be separated by dissolving the metal in dilute hydrochloric acid, adding enough bromine water to keep the solution well oxidized, and to prevent loss as silicuretted hydrogen. (Boron, if added to aluminum, is found to exist similarly in the metal in the dual form of graphitic and diamantine boron.-Deville.)

The following table gives the average composition of the metal of the various makers, of given percentages of contained aluminum, as shown in many analyses made by the Pittsburgh Testing

tory:

TABLE I.

Composition of Various Commercial Aluminums.

Labora

1.35 2.00

1.00

0.04

0.03

2.00 0.01 0.07 uil.
2.80 1.49

95.00 1.50
0.07
95.00 0.90
3.00 0.05 0.01 0.04
95.00 0.90 0.75 3.25 0.02 0.07 0.01
95.00 0.85 0.75 3.00 0.10 0.30 nil.
95.00 1.75 1.15
94.15 1.02 0.54
97.00 1.55 1.25

0.13 0.03 0.02 0.02
97.00 0.90 0.82 1.20 0.04 0.01 0.03
97.00 0.95 0.53 1.45 0.01 0.05 0.01
97.00 0.75 0.52 1.55 0.03 0.15 nil.
98.00 1.30 0.60 0.07 0.01 0.01 0.01
98.00 0.71 0.35 0.90 0.01 0.01 0.02
98.00 0.90 0.29 0.75 0.01 0.04 0.01
98.00 0.90 0.23 0.80 0.02 0.05 mil.
98.52 0.42 0.72 0.05 0.06 nil. 0.04
99.00 0,80 0.15 0.03 0.01 nil. 0.01
99.00 0.35 0.13 0.50 nil. 0.01 0.01
99.00 0.31 0.20 0.45 nil. 0.03 0.01
99.00 0.27 0.15 0.55 0.02 0.01
99.20 0.41 0.34 0.05 nil. nil.
99.14 0.23 0.17 0.46 nil. nil.
99.34 0.40 0.21 0.05 nil. nil.

nil. nil.

nil.

nil.

A sample lot of aluminum made by the Pittsburgh Reduction Company, taken to represent as fairly as possible the average metal made by this company, was submitted to many physical tests, which are reported in various parts of this paper. For the purpose of identification and brevity, this metal will be referred to as "the average lot 98.52 per cent. aluminum."

This metal had the following composition:

PROPERTIES WITH REFERENCE TO SPECIFIC GRAVITY.

One of the most striking properties of aluminum is its lightness, which, for many purposes, makes it comparable in value with metals costing one-fourth as much.

The following table gives the specific gravities of aluminum, with authorities:

The specific gravity of aluminum being taken as one, soft steel i

very nearly 2.95 times, copper 3.6 times, ordinary high brass 3.45 times, nickel 3.5 times, silver 4 times, lead 4.8 times, and gold 7.7 times as heavy.

A sheet of aluminum 12 inches square and 1 inch thick weighs 14.03 pounds; a bar of aluminum 1 inch square and 12 inches long, 1.17 pounds; a bar of aluminum 1 inch in diameter and 12 inches long, 0.918 pounds.

Pounds.

Aluminum has about the tensile strength of cast-iron, with only about one-third of its weight, and may be cast as easily and successfully. It will therefore be advantageously used to replace castiron in the parts of moving machinery that have to be reversed or otherwise have their momentum overcome; since, at one-third the weight, they will have only one-third of the momentum.

ACTION OF HEAT.

Pure aluminum melts and becomes fluid at about 1200 degrees Fahrenheit. The amount of impurity in aluminum materially alters its melting-point. One per cent. of iron raises the melting-point over 100 degrees: It does not, like lead, remain firm almost to the fluid-point and then suddenly give way, but passes through a stage from 1000 to 1200 degrees Fahrenheit, in which the metal becomes pasty, loses much of its power of cohesion, and, if it be gently pressed together, can be readily welded. It is, however, very red-short at this temperature, and will not stand hammering to weld the metal without crumbling down. If the metal is not held too long a time in this pasty condition, it does not seem to have been injured after being again cooled down.

The Thomson Electric Welding Company has no difficulty in welding aluminum satisfactorily, rapidly and cheaply, and so firmly that the welded strips will break in the body of the bar. This is done in the admirable Thomson machine, by which the surfaces to be welded are brought together at the welding-temperature (near that of fusion, for many metals, and a temperature at which the rigidity is almost entirely lost in all of them) automatically, in such a way that the welded faces will cohere without changing the form of the constituent parts.

The melting-point of aluminum, as given by Roscoe & Schor