272 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1910. 



against a friction clutch in the ordinary way during the stop. At 

 the end of the fifth second each cylinder is automatically released by 

 chronometric means, in the manner shown in the next diagram 

 (fig. 8). 



Here 3*011 will see that a special form of clock is used, with a center 

 seconds' hand which projects beyond the face by about an inch, and to 

 the end of it is attached a brush of exceedingly fine silver wires. At 

 every twelfth part of the circumference of the clock dial is fixed a 

 platinum pin, and consequently every five seconds the little brush 

 wipes against the convex surface of one of them. Each of these pins 

 is connected with one terminal of a battery B, the other side of the 

 battery leading to the relay R, as does also the center seconds' hand. 

 Therefore each time the brush wipes against a pin the circuit is closed, 

 and the relay throws into action the local circuit connected up with 

 the terminals TT. This circuit excites an electromagnet, which 

 attracts an armature and pulls away the check which is holding 

 back the cylinder. At the end of each five seconds the cylinders 

 consequently recommence turning. 



Well-calibrated clocks of the pattern used will keep good time for 

 the period taken to transmit a picture, one gaining on the other quite 

 an inappreciable amount, depending on the friction of the brush 

 against the pins. By this means the two cylinders are kept in very 

 fair synchronism independently of any wireless communication, and 

 the less the interval between the stopping and restarting of the cylin- 

 ders be made, the more accurate and satisfactory will be the effect. 



The other method of synchronizing is controlled by electromag- 

 netic oscillations. Let us suppose that a coherer is being used as 

 cymoscope; the transmitting cylinder is kept running without any 

 interruption, but by means of a fleeting contact it sends out a wave 

 at the conclusion of its turn, a bare space in the picture being neces- 

 sary about half a second beforehand, so that no waves are sent out 

 for the half-second previously. The receiving cylinder is driven too 

 quickly, and checked at the end of the revolution. It then, by means 

 of a cam pressing down a spring lever, throws out of circuit the 

 marking current, and brings into circuit the relay which actuates 

 the electromagnetic release. Consequently, when the synchronizing 

 wave is received, the coherer causes the relay to work, the release is 

 effected, and the receiving cylinder starts a new revolution in unison 

 with the transmitter. 



This means of synchronizing is only possible in cases where a 

 cymoscope is employed that is capable of actuating a relay, and you 

 will therefore see that it is out of the question, except for short dis- 

 tances. I am therefore using the chronometric system already de- 

 scribed in the apparatus, and it is being embodied in the quartz fiber 

 apparatus I am now about to describe. I must first remark that the 



