396 



PROCEEDINGS OF THE AMERICAN ACADEMY.. 



Dark 

 ■Room 



Figure VI. 



of the sensitive plate is covered, so that the observer, who charges the 

 condenser from a step-up transformer by operating a switch in the dark 



room, can see a flash upon 

 the pLate when the plate 

 is struck. In fact, he can 

 see quite plainly each of 

 the oscillations, though 

 they persist in some cases 

 for only a fraction of a 

 millionth of a second. 



In order to measure the 

 time of the discharge, the 

 speed of the mirror was 

 obtained by the aid of a 

 stroboscopic device (Fig- 

 ure VII) as follows : A 

 small aluminum disc D, 

 marked off in alternate black and white sectors, is attached to the axis 

 of the mirror. The disc is illuminated periodically by flashes in a Geiss- 

 ler tube G, connected to the secondary of an induction coil C, of which 

 the primary is interrupted by an electrically driven tuning-fork T. The 

 tuning-fork makes 256 vibrations per second. The disc contains 12 black 

 sectors ; so that if the disc makes i^g of a revolution between two consec- 

 utive flashes in the tube, the disc will appear to stand still. Thus by 

 observing the disc (by a telescope through the wall of the dark room) 

 and adjusting the resistance in the field of the motor that drives the 

 mirror, the disc is brought to an apparently stationary condition. The 

 mirror is then making ^^^ of 256 = 21.33 revolutions per second. Other 

 apparently stationary conditions of the disc correspond to 42.66 and 64 

 revolutions per second. It is not difficult to set on these speeds with an 

 accuracy of one or two tenths of one per cent. 



The mercury interrupters of which the accompanying photographs 

 were taken were so constructed that the mercury surfaces were brought 

 near to each other (about 1 mm.), so that the image of both electrodes 

 fell near together on the plate. 



The revolving-mirror photographs of the mercury interrupter and of 

 the ordinary spark in air between cadmium terminals are shown in 

 Plate I. 



AVhen the mercury interrupter in action is viewed directly by the eye 

 at rest, without the intermediation of the revolving mirror, it shows 



