March 10, 1916] 



SCIENCE 



355 



The chief obstacle in employing a free- 

 swinging pendulum for graphic purposes is the 

 difficulty of determining its rate. The meth- 

 ods of coincidence devised by physicists neces- 

 sitate the use of a standard clock, and are 

 somewhat complicated. The writer has con- 

 structed a simple apparatus for recording the 

 exact number of beats in the period tested, 

 which has given very satisfactory results. The 

 figure shows the essential details of its con- 

 struction. 



The common base of a telegraphic sounder 

 and key is fastened to a larger board. A 

 counter for registering rotation is screwed to 

 a small block, with its axis at right angles to 

 the lever of the sounder. A brass disc with 60 

 ratchet teeth is attached to the end of the axis. 

 The movements of the lever of the sounder are 



communicated to the toothed wheel by two 

 strips of steel which are connected by a joint. 

 The first strip is fastened to the right-hand 

 end of the lever with a screw. A pin projects 

 at right angles from the lower end of the sec- 

 ond strip. This pin is shaped to fill the inter- 

 val between two teeth, and acts both as a pawl 

 and an index. The front face of the disc is 

 divided by radial lines which mark the posi- 

 tion of every fifth tooth, the tenths being 

 designated by longer lines. 



The sounder is placed in the circuit, which 

 is periodically closed and opened by the pen- 

 dulum that is being tested. When the lever is 

 pulled down by the electromagnets the disc is 

 rotated to the extent of one tooth, and every 

 complete rotation of the wheel is registered by 

 the counter. To prevent the disc from ad- 

 vancing too far when it is suddenly pulled 

 forward, and from recoiling when the pawl 

 moves backward, a clock spring, fastened to a 

 second block, is pressed against the back of 

 the wheel by a screw with a milled head pass- 



ing through the block. This screw does not 

 appear in the sketch. The degree of pressiu'e 

 required by varying conditions is regulated by 

 experiment. The index is pressed against the 

 wheel by a delicate spiral spring stretched 

 across the angle of the arm. The extent of 

 movement of the arm is regulated by the two 

 screws at the left of the figure. The dropping 

 of the pawl is due to the spring attached to the 

 end of the lever, the tension of the spring 

 being regulated by the screw to which its lower 

 end is fastened. Pieces of paper are placed 

 between the armature and the magnets in 

 order that the lever may return with sufficient 

 rapidity to its original position. When it is 

 desired to use the sounder without the counter, 

 the screw which clamps the arm to the lever is 

 loosened, the arm elevated a little, and the 

 screw again tightened. 



The electric counter should be used with a 

 reliable time-piece indicating seconds. The 

 zero mark on the disc is set opposite the index- 

 pawl. When the second hand of the watch is 

 in the desired position, the horizontal lever to 

 which the right hand knob is attached is thrust 

 beneath the spring, which closes the circuit. 

 While the instrument is in operation, the in- 

 stant when the zero notch is closed by the 

 index is noted from time to time. If at the 

 end of an hour no variation between the indi- 

 cations of the disc and the watch can be 

 detected, the seconds pendulum is considered 

 accurate enough for testing purposes. It 

 should have an error less than one half second 

 in an hour. In using this coincidence method 

 in regulating the rate of the pendulum it is 

 not necessary to read the counter. The 

 counter is only indispensable in finding the 

 number of vibrations when the rate is un- 

 known. 



The seconds pendulum which I have em- 

 ployed in my experiments oscillates upon knife- 

 edges of hardened steel, and has a bob weigh- 

 ing 2.4 kilograms. A platinum wire in the 

 lower end of the rod makes contact periodically 

 with a globule of mercury. The length of the 

 tangent of one degree of arc described by the 

 wire in the end of the rod is two centimeters. 

 By holding a millimeter scale horizontally by 



