WIRELESS TIME SIGNALS. 



271 



the signals coincide with the beats of our A (or B) clock. 

 The principal difficulty seems to be this — that a very short 

 sharp "dot" cannot be radiated and received over long 

 distances, and that if the dot is of an appreciable length 

 different observers take different parts of it to form their 

 coincidences with their clock beats. The personal element 

 also arises from the fact that the sound of the radio signal 

 generally differs greatly from the sound of the clock beat, 

 however it is introduced. We propose to avoid both these 

 difficulties and eliminate the personal element by means of 

 the following simple mechanical device. 



The clock circuit, 

 as usual, passes from 

 the clock contacts 

 through a battery to 

 the relay R, and for 

 all practical purposes 

 R is the clock. The 

 armature of the relay 

 is pivoted at B and 

 normally rests against 

 the stop at A. When 

 clocK the seconds-impulse 



comes from the clock the contact A is broken and the 

 armature rests against the stop C. Let us now interpose 

 the contact A in the path of the incoming signals. That 

 is, let D A B E be in series with the aerial. That is all. 

 Let us follow the sequence, supposing that the period of 

 the radio dots is 0*99s, thus giving a coincidence, with a 

 gain of one second upon the local clock, in every hundred 

 seconds. 



With relay open, as shown in the above diagram, the 

 radio-dots come through uninterruptedly, but with relay 

 closed, the radio circuit is broken and we hear nothing. For 



