which hangs a scale-pan, d. The ratio of c, to c, is 500: 10. The central fulcrum of the lever rests on the end of the ram, b, so that the whole measuring apparatus moves along as the piece extends

and the ram moves out, the arm, c,, being always kept horizontal by the aid of a spirit level.

Wicksteed's machine* (Fig. 5) is a vertical one with a single lever, c, c2, placed horizontally on the top. A movable poise, m, measures the load, pressure being applied to the ram, b, by a screw or pump. The ratio of c, to c, is 50: 1. In a 100-ton machine, the weight, m, is I ton, so that it balances a pull of 50 tons when at the end of c. To carry the load on to 100 tons, m is run back beyond the fulcrum, and a second weight of 1 ton is hung to the end of c. The poise weighs 1 ton, ard is moved along the lever by a screw worked by power. Each 3 inches of movement of the poise adds I ton to the load on the test-piece, whilst a vernier

*Inst. Mech. Eng. Proc., 1882, p. 384.

attached to the poise may be read, on a scale affixed to the lever, to one-hundredth of a ton.

Greenwood's machine (Fig. 6) is in use in the laboratory of Professor Kennedy, by whom the original type has been con

siderably improved. It is a horizontal machine with two levers, a knee lever, ff (5: 1), and a steel-yard, c, c, (201), the total leverage being 100: I. The load is applied by the ram, b, and measured by the position of the poise, m, on the steel-yard.

Gollner's machine (Fig. 7) is a double-lever vertical machine,

working up to 20 tons. Both screw and ram may be provided, with means for changing at once from one to the other. For light tests, the lever, fif,, can be disconnected, and the machine used as a single lever machine. This type of testing machine is that used in the laboratory of the Royal School of Mines, and a

In a testing machine it is extremely desirable to have some apparatus by the aid of which the stress-strain diagram of a piece of metal under test may be drawn automatically. The most successful apparatus of this kind is that devised by Wicksteed*

Inst. Mech. Eng. Proc., 1886, p. 27. For descriptions of American machines consult Testing Machines: their History, Construction and Use, by A. V. Abbott. New York, 1884.

(Fig. 9). In this, the motion of the pencil that indicates the load is derived from the pressure in the hydraulic press, and not from the weighing apparatus, a wire attached by clips to the specimen serving to rotate a recording drum by an amount proportional to the elongation. The pencil having an axial motion proportional

α

FIG. 9.

to the load, and the drum a rotating motion proportional to the extension, a stress-strain diagram is described.

The following table gives the ultimate tensile strength of a number of metals in lbs. avoirdupois per square inch :—

Influence of Foreign Elements on the Strength of Metals. The influence of chemical composition on the mechanical properties of metallic masses is of great importance; this has long been recognised, but it is singular that the subject has been so little investigated. Turn, for instance, to Dr. Percy's classical work on Iron and Steel, published in 1864. It fully represents the information which had been gained at that time, yet it con

tains the results of but few mechanical tests on the materials with which it deals.

*

The influence of foreign elements are best shown in the case of iron. The properties of this metal are absolutely changed by the presence of a few tenths per cent. of carbon; but it may be doubted whether the exact influence exerted by varying proportions of carbon has been accurately determined. A distinguished authority, Mr. H. M. Howe, thinks that it is not yet known, for, in an elaborate work recently published on the Metallurgy of Steel (New York, 1890), which will be of much service to the student, he points out that he has plotted, in a single curve, the results of over 2500 tests, and yet the conclusion he arrives at is, that we are not at present able to quantitatively express the effect of carbon. The fact is, that metallurgists are only beginning to realise that the effect of elements in the presence of each other is very complicated, and that it is absolutely necessary to study the effect of any given element on an absolutely pure mass of the metal to be tested. With regard to iron, the author has shown that the tensile strength of electro-deposited iron, which is as pure as any iron can be, is 2.7 tons per square inch before annealing, and 15.5 tons per square inch after annealing. Even in this case, however, it is doubtful how far the result is influenced by the presence of occluded hydrogen, or by the fact that electrolytic iron is probably an allotropic form of the metal. In studying the effect of carbon, it is very difficult, therefore, to start from pure iron as an absolute basis. It may be sufficient, however, for the student to keep steadily in mind the very useful diagram, represented in Fig. 10, by which Deshayes † has indicated the effect of carbon, and incidentally that of manganese, on the tensile strength of iron. The co-ordinates are respectively percentages of carbon and tensile strengths expressed in tons per square inch. The effect of carbon is indicated by the dark continuous line, and that of varying proportions of manganese in the presence of carbon by the thin continuous, and the dotted lines. The effect of a given proportion of carbon is found by producing a vertical line, from the point indicating the percentage of carbon, until it cuts the darker curve, so that the tensile strength will be represented by the length of this vertical line. It may be added that, according to some authorities, the maximum strength would be attained in steel containing 0.8 to 1.0 per cent. of carbon, after which point the line representing tensile * Journ. Iron and Steel Inst., No. i. (1887), p. 74. + Classement et emploi des Aciers (Paris, 1880), p. 151.