Mendenhall — Free Pendulum as a Time Standard. 87 



plane passing through the knife edge. A similar mirror is 

 placed parallel to this and as near to it as is possible without 

 interfering with the motion of the pendulum, but it is rigidly 

 attached to the support upon which the pendulum swings. 

 What is called the u flash apparatus" is placed a meter or two 

 meters in front of these mirrors and in a line normal to them. 

 In this an electric spark may be used or the light of a lamp or ■ 

 candle. What is desired is that a flash of light should be 

 produced every second, as determined by a break-circuit 

 chronometer or clock. For use with a lamp Mr. E. Gr. Fischer 

 has contrived an ingenious shutter operated by an electro- 

 magnet, by means of which an illuminated slit is exposed for 

 an instant in the movement following the break only. An 

 induction coil whose primary circuit is broken by an electro- 

 magnet in the chronometer circuit provides the electric flash. 

 Just over the point where this flash is produced is placed a 

 telescope. 



When properly adjusted the flash is reflected from both 

 mirrors, and assuming the pendulum to be at rest, two lines of 

 light are seen as one in the telescope or better, one is made to 

 overlap the other a little. Now suppose the pendulum to be 

 in motion, it will be clear that, as the flash occurs only for 

 an instant each second, whether it will be" seen reflected from 

 the moving mirror will depend on the position of that mirror 

 at the instant of its appearance. If it should happen to be in 

 the plane of its original adjustment when the flash occurs the 

 appearance in the telescope will be precisely the same as when 

 the pendulum is at rest. If the period of the pendulum be 

 precisely one-half of that of the clock or chronometer it will 

 return to this position in just one second and the appearance 

 will be continually repeated. If, however, the pendulum be 

 slightly slow or fast of the chronometer, the mirror will not 

 be precisely in this position at the end of one second and the 

 image from its mirror will be a little above or below that of 

 the first mirror. In another second the distance separating 

 them will be still greater and this will go on until the moving 

 image is no longer seen in the field of the telescope. After a 

 time, however (this period is conveniently about five minutes), 

 the pendulum will have gained or lost one oscillation on the 

 beat of the chronometer and a few seconds before the neces- 

 sary period for this has elapsed the image re-appears in the 

 field and approaches coincidence, to again recede on the other 

 side. 



It is only necessary to observe the instant of this coinci- 

 dence of the two images. Under the conditions mentioned 

 above it is not necessary to observe this closer than the nearest 



