420 STRENGTH OF MATERIALS by a hand lever, is used to force oil into the press. The breaking force is thus applied from below, and is measured upon the lever above. 26. With the autographic recording testing machine of Prof. R. H. Thurston, fig. 2, near- ly all of the essential qualities as well as the strength of ma- terials are deter- mined by the au- tomatic production of a strain diagram. This diagram is an exact graphical rep- resentation of all circumstances at- tending the distor- tion and fracture of the specimen. No system of personal observation yields results as trustwor- thy or with such precision as an au- tographic registry. No other method gives simultaneous- ly, and at every instant during the test, the intensity of the distorting force and the magnitude of the coincident distortion. In this machine two strong wrenches are carried by the A frames, and depend from axes which are both in the same line, but which are not con- nected with each other. The arm of one of these wrenches carries a weight at its lower end. The other arm is designed to be moved by hand in the smaller machines and by a worm gear in larger ones. The heads of the wrenches are fitted to take the head on the end of the test Sieces, which are usually given the form shown i fig. 3. A guide curve of such form that its FIG. 2. Thurston's Testing Machine. FIG. 8. Test Piece. ordinates are precisely proportional to the tor- sional moments exerted by the weighted arm while moving up an arc to which the corre- sponding abscissas of the curve are propor- tional, is secured to the frame next the weight- ed arm. The pencil holder is carried on this arm, and as the latter is forced out of the vertical position, the pencil is pushed forward by the guide curve, its movement being thus made proportional to the force which, trans- mitted through the test piece, produces deflec- tion of the weighted arm. The guide curve is a curve of sines. The other arm carries the cylinder upon which the paper receiving the record is clamped, and the pencil makes its mark on the table thus provided. This table having a motion, relatively to the pencil, which is precisely the angular relative motion of the two extremities of the test piece, the curve de- scribed upon the paper is always of such form that the abscissa of any point measures the amount of the distortion which the force pro- duces. 27. The vertical scale of the diagrams produced is a scale of torsional moments, and the horizontal scale is one of total angles of torsion. Since the resistance to shearing, in a homogeneous material, varies with the resis- tance to longitudinal stress, the vertical scale is also for such materials a scale of direct re- sistance; and with approximately homogene- ous substances this scale is approximately ac- curate, where, as here, all specimens compared are of the same dimensions. 28. By fig. 4 it will be seen that the first portion of the line rises at a slight inclination from the vertical, and very nearly straight. The amount of dis- tortion here is seen to be approximately pro- portional to the distorting force', illustrating Hooke's law, Ut tensio sic vis. After a degree of distortion which is determined by the spe- cific character of each piece, the line becomes curved, the change of form having a rate of increase which varies more rapidly than the applied force. When this change begins, the molecules, which up to that point retain gen- erally their original distribution, while varying their relative distances, begin to change their positions with respect to each other, moving upon each other in a manner similar to that action described by H. Tresca, and called the u flow of solids." This point, at which the line begins to become concave toward the base, is considered as marking the torsional limit of elasticity. It is well defined in experiments upon woods ; is less marked, but still well de- fined, in the fibrous irons and the less homoge- neous specimens of other metals ; and becomes quite indeterminable with the most homoge- neous materials, as with the best qualities of well worked cast steel. This point does not indicate the first set, since a set occurs with every degree of distortion, however small. It is at this elastic limit that the sets begin to be- come proportional to the degree of distortion. The inclination of the straight portion of the line from the vertical measures the stiffness of the specimen. This rigidity is very closely, if not precisely, proportional to the hardness, in homogeneous substances ; and this quan- tity is taken, for practical purposes, as a mea- sure of the hardness of the metals and of their elastic resistance to compression. After pass- ing the elastic limit, the line becomes more and more nearly parallel to the base line, and then, with the woods invariably, and in some cases with the metals, begins to fall before fracture becomes evident in the specimen. With the more ductile substances, nearly all the particles are brought up to a maximum in resistance before fracture occurs, and this cir- cumstance has an important influence in deter-
