26 THE STUDY OF SPEECH CUKVES. 



lever by a connecting link. Since it is desirable to have considerable 

 weight to keep the tracing point at the bottom of the groove, the first 

 lever can be made of a stout brass rod. The connections are made by 

 gimbal joints. 



The gimbal joint (figure 23) comprises a steel square at whose corners 

 pivot holes are bored. Each fork carries a fixed and an adjustable pivot 

 screw, the adjustable one having a jam nut. It is vitally important to 

 have the axial lines correct and all pivots and pivot holes accurately round 

 and central. The rod of one fork is inserted into a hole in the end of 

 the first lever, that of the other fork into a hole in the connecting link. 

 The connecting link is a tube of aluminum. The connecting link moves 

 the second lever by a joint (figure 24) whose construction is the same 

 as that of the gimbal joint, except that one fork carries a metal piece with 

 a hole for the second lever. 



The second lever must be light and stiff. The central portion is a 

 piece of aluminum tul^ing; the other portion is a reed or stiff straw. The 

 degree of amplification depends on the length of the lever and the nearness 

 of the link to the fulcrum. The attachment to the support is by a fork 

 with pivots working on an axle. 



It would naturally be expected that the magnification of the speech 

 vibrations could be carried to any desired extent l)y means of the com- 

 pound lever. The difficulty lies in the making of the pivots at the joints 

 and in adjusting them. The cones of the pivots and the centers of the 

 bearings should be perfectly true to the desired axes; the joints should 

 be perfectly tight and yet perfectly frictionless. The finest mechanics 

 obtainable for making the joints have not succeeded in producing a com- 

 pound lever that correctly enlarges more than 125 times. The best results 

 have been obtained with the single lever with which correct tracings with 

 a magnification of 300 times have been reached. The future of the 

 method lies, however, in the development of a compound lever. 



The recording point (figure 25) consists of a piece of thin card cut 

 across and joined by the thinnest obtainable non-elastic membrane (gold- 

 beater's skin); one piece is fastened to the lever, the other carries a fine 

 glass rod with its end rounded by heat. This recorder is stiff in the direc- 

 tion of the movement, but highly flexible at the hinge. The critical part 

 lies in the meml^rane; this must be so thin that its elasticity has no effect 

 in pressing the glass point against the drum. The recording point is so 

 adjusted that the minute glass knob is in a horizontal plane passing through 

 the axis of the drum, and the hinge is almost but not quite under the 

 knob (figure 26). If the angle of the glass rod with the surface of the 

 drum is too great, the point will catch and move by jerks. The axis of 



