44 THE ROYAL SOCIETY OF CANADA 



considered sufficient for accurate results. It was therefore decided 

 to attempt to make use of the Canadian station at Kingston for the 

 transmission of such signals as were required ; the only point in doubt 

 was whether the range of the station was sufficient to reach the points 

 required. Upon further experiment this unfortunately turned out 

 not to be the case, and it was necessary to fall back upon Arlington. 



Since the Arlington signals are controlled by a mean time clock 

 it is not feasible to make a direct comparison with a sidereal chronom- 

 eter by the method of coincidence of beats, since it would frequently 

 happen that a coincidence would not occur during the five minutes 

 for which the signals continue. It was therefore necessary to intro- 

 duce an intermediary chronometer having a large rate; at the outset 

 this was rated to gain ten minutes per day on mean time ; the difference 

 of rate was later increased to fifteen minutes per day, giving a coin- 

 cidence every 96 seconds approximately; three coincidences during 

 the five minutes were thus assured, and owing to their spacing it was 

 impossible that more than one of these should fall in the silent space at 

 the end of the minute. 



To make the comparison, a circuit derived from the chronom- 

 eter, and including a suitable inductance, is led close to the cord 

 attached to the head-piece of the receiving set; when properly ad- 

 justed this causes a sharp click in the telephone on the opening of the 

 chronometer circuit, while the subsequent closing of the circuit is 

 almost inaudible. Owing to the difference in rate, these chronom- 

 eter clicks gradually gain on and pass through the wireless beats; 

 the beat when the click coincides with the beginning of the wireless 

 signal is the moment of coincidence. The chronometer times of the 

 coincidences having been noted, as well as the time (to the nearest 

 half second) of one or more Arlington even minutes, the compaiison 

 between Arlington and the chronometer is established ; this is bridged 

 over to the sidereal observing chronometer by a chronographic com- 

 parison before and after the receipt of signals. 



We have now a comparison, at a particular instant, between 

 Arlington time and the observing clock at one station. It is obvious 

 that if a similar comparison has been made at another station, we 

 can deduce the difference of time by the two clocks, which completes 

 the longitude "exchange." All the other operations incidental to a 

 determination of longitude are carried on in the usual way. 



It is to be noted that the deduced longitude does not depend at 

 all on the correctness of the time signals used, so long as their rate is 

 correct; thus any other set of rhythmic signals would serve the purpose 



