Section III, 1918 [43] Trans. R.S.C. 



The Application of Wireless by the Dominion Observatory for Longitude 



Determinations 



By R. M. Stewart, M.A. 



Presented by Otto Klotz, LL.D., F.R.S.C. 



(Read May Meeting, 1918.) 



As no account has been given of the appHcation of wireless 

 telegraphy to the determination of longitudes in Canada, it seemed 

 that a short description of the work done and the methods used by 

 the Dominion Observatory might be of interest. 



In the telegraphic method, the function of the telegraph Hne is 

 the making of a direct comparison between the observing clocks at 

 the two stations; in practice this is done by connecting the telegraph 

 line to the chronograph at each station and sending a certain number 

 of arbitrary signals in each direction over the line; the signals are 

 recorded on both chronographs, together with the beats of the respec- 

 tive clocks; this affords the comparison desired. 



When it is desired to use wireless telegraphy a different procedure 

 is required, since wireless signals cannot readily be recorded on a 

 chronograph. It is hence necessary to resort to the method of co- 

 incidence of beats, a method of comparison which may be made very 

 nearly, if not quite, as accurate as a comparison by chronograph. 



When the necessity arose in the spring of 1914 for the determina- 

 tion of longitudes of several stations in Quebec which were inaccessible 

 by telegraph, it was decided to inaugurate the use of wireless teleg- 

 raphy for the purpose. Two courses were then open. It was possible 

 to make use of the wireless time-signals sent out by the U.S. naval 

 station at ArHngton, near Washington, for distribution of time at 

 sea; on the other hand, special signals might be sent out when re- 

 quired from one or other of the Canadian Government stations. The 

 Arlington signals are sent out every night from 9.55 to 10; they 

 are controlled by a clock beating seconds, the 29th second and the 

 last five seconds of each minute being omitted for identification ; the 

 duration of individual beats is 3^ second or less. 



Preliminary experiments had shown that beats of about }4. 

 second duration, with only one beat per minute omitted, were pre- 

 ferable for our purpose; also the time available (5 minutes) was not 



