64 THE TRANSATLANTIC LONGITUDE. 



IX. 



PERSONAL ERROR IN NOTING SIGNALS. 



Since the signals sent through the telegraphic cable were recorded upon the 

 chronograph automatically at the transmitting station, but at the receiving station 

 through the mediation of an observer, who noted the deflection of the light-spot 

 from the galvanometer by sending a second telegraphic signal to his own chrono- 

 graph, it will be seen that the interval a;, which elapses between the giving of a 

 signal at one station and its chronographic record at the other, may be conve- 

 niently divided into four diff"erent parts, viz., the time requisite 



1. For the signal to arrive at the other station; 



2. For the magnet of the galvanometer to be moved through an arc sufficient to 



be readily perceived ; 



3. For the observer to take cognizance of the deflection, and give his signal 



upon the break-circuit key ; 



4. For this observation-signal to be recorded upon the chronograph. 



Each of these four parts comprises the time, appreciable or otherwise, consumed 

 in more than one distinct process ; yet this division suffices for all our purposes. 

 If these several intervals be practically equal at the two stations, they become 

 absolutely eliminated in our determination of the longitude. If they be unequal, 

 the resultant longitude will require an increase by one-half the excess of their sum 

 for westward signals. In either case, only their total sum at the two stations is 

 determined by the operations for longitude. 



If we assume that the time lost upon the chronograph-circuit is the same at each 

 station, the last of the above-mentioned intervals becomes eliminated by the com- 

 parison of the two records. The second and third depend upon the galvanometer 

 and observer at the receiving station, and are not easily to be separated from each 

 other in any determination of their amount ; but if their sum can be measured, 

 this, subtracted from our quantity tc, will aff"ord a trustworthy determination of the 

 velocity with which the signals are actually transmitted through the telegraphic 

 circuit. 



This sum of the delays dependent on the galvanometer and the observer, I have 

 called "the personal error of noting;" and the attempts to measure its amount have 

 been so successful, and have manifested such an unexpected constancy in its value 

 for different persons, at different times, and at both stations, that the results 

 obtained for the velocity of transmission of our signals seem entitled to a high 

 degree of confldence. 



By observing a series of signals similar to those exchanged for longitude, and so 

 arranged that both the original signal and the observation of the consequent 

 deflection shall be recorded on the same chronograph, the desired measure may be 

 obtained. Experiment showed at once that the interval thus determined was alto- 

 gether too large for any inconvenience to arise from the use of a single recording 

 pen. The obstacle first encountered arose from the circumstance that the minimum 

 battery force requisite for the electro-magnet of the chronograph pen Avas about 

 seventy-five times greater than the maximum which could be safely employed for 



