150 PROCEEDINGS OF THE AMERICAN ACADEMY' 



light friction in the narrow groove between the two disks as the pen- 

 dulum swings ; both the end of the wire and the sharp edges of the 

 disks are somewhat rounded to prevent undue wear. If the two disks 

 are clamped together in a position such that the unarmored part of 

 one is opposite the armored part of the other, and that as the pendulum 

 swings the wire leaves one German silver strip just as it reaches the 

 other, it will be in contact with each armature successively, but will 

 establish metallic contact betweeu them and close the circuit between 

 the mercury cups at most for a very short interval ; if, however, the 

 armatures overlap for a certain number of degrees, contact will be 

 maintained while the pendulum swings through just that number of 

 degrees of its arc. The front disk, which is rigidly fixed to the pen- 

 dulum rod, is so placed that it comes in contact with the copper wire 

 just as the pendulum reaches its lowest point. If therefore the arma- 

 tures overlap, for instance, 10°, and if the pendulum is urawn out to 

 a known point upon one side of its arc so that the copper wire rests 

 upon the unarmored side of the " fixed " disk, then, when the pendu- 

 lum is allowed to fall, metallic connection will exist between the disks 

 while it is rising through 10° from its lowest point. From this 

 "arc of contact" the time of contact is easily determined. The time 

 of vibration of the pendulum is 1.181 sec, and the arc of initial 

 displacement 28° 56'; the decrease of arc in one complete vibration 

 is about 30'. 



The length of the arc of contact may easily be determined within 

 one minute. This uncertainty may give rise to an error in the com- 

 puted time varying from 0.00022 sec. for an arc of 1°, to 0.00042 sec. 

 for an arc of 25°. 



The errors arising from a wrongly assumed value of the initial 

 displacement, and from decrease of arc, are in no case greater than 

 one fourth of one per cent of the computed time, and are here 

 neglected. 



Since the copper wire does not touch each of the armatures in 

 a single point, the arc of contact will obviously not be exactly equal 

 to the angle of overlap, but its value may be easily determined by 

 direct observation. By throwing the switch at H to the right, the 

 circuit through the bell and battery, J, is made complete, except for 

 the break at the disks. The pendulum is then lowered, and the 

 readings on the divided arc E of the microscope K are noted when 

 the bell begins and stops ringing, as the pendulum is moved through 

 its ascending arc ; the former must nearly correspond with the 

 reading when the pendulum is in its position of rest. To facilitate 



