728 Experiments on the communication of [Sept. 



Although the watches were of the very cheapest kind, and would 

 not keep time together for more than five minutes, still they were 

 quite sufficient to enable a correspondence to be carried on. Thus a 

 signal seldom lasted longer than three minutes ; both watches were 

 then allowed to run to No. 1 or zero, and stopped. To renew corres- 

 pondence a prolonged roll was communicated. If but one roll, it in- 

 dicated spelling ; if two, numbering. On the roll ceasing, three pulsa- 

 tions at intervals of one second were passed, and at the third the 

 correspondents started their watches. 



The pendulum was also tried, and with decided advantage. Two 

 German clocks sufficed to demonstrate the practicability of the system. 

 The striking parts were removed, and also the hour and minute hands 

 and dial. — To the axis of the escapement wheel a needle was attached, 

 carrying a light hand which indicated on a dial the signals above 

 described. The German clocks (which cost but 16 rupees the pair,) 

 in numerous experiments beat together for several hours, and could 

 always be relied on for one hour at least. It is almost needless to add, 

 that by shortening or lengthening the pendulum the rate was readily 

 varied from 40 to 80 seconds for each revolution. 



I did not omit chronometers, although I could not of course so alter 

 these costly instruments as to adjust them perfectly to my experiments. 

 It is obvious however that chronometers will on my method give 

 an unerring and constant mode of telegraphic correspondence. In a 

 recent trial at Greenwich the mean error of several instruments in 

 one year was but two seconds!* Here then are two movers con- 

 stantly and simultaneously pointing to one and the same signal, be it 

 letter, figure, or cypher. The electric pulsations which " take no note 

 of time" or distance, supply us with the means of converting this 

 synchronism to the unexpected and invaluable end to which it is now 

 proposed to be applied. 



Even employing inferior chronometers, the addition of a moveable 

 dial which could be adjusted daily on a method too simple to need 

 description, would secure the perfection of the correspondence ; or the 

 daily difference of the instruments being known, a tabular correction 

 could we made ; or, lastly, by an occasional astronomical observation 

 of true time at each station, the object in view could be as certainly 

 obtained. 



* In 1831, the first three prize chronometers only differed -J of a second in one 



100 

 year. 



