THE TRANSATLANTIC LONGITUDE. 



93 



nected at the sending station during the intervals between the signals, but the bat- 

 tery was short-circuited there also. Thus the cables were always resuming their 

 equilibrium between the signals, during each of the five nights when the exchanges 

 for longitude were made ; there being upon the first two nights only one length of 

 cable used, but upon the last three a double length, through Avhich the adjustment 

 of the perturbation was to be effected. 



I will give the results for these five nights in the same form in which they were 

 first presented ; viz., the mean difference between the records of the same signals 

 upon the two registers, this being the resultant value of the longitude, uncorrected for 

 clock-errors or for transmission-time. The 2h. 52m., which are common to all, 

 can be here omitted, only the seconds and fractions of a second being needful for 

 our purpose, and the signals are assorted according to the length of the interval 

 which immediately preceded. On each date three series, of 20 signals each, were 

 sent from each station, but not all were received. The average number upon which 

 the several values for each day actually depend, is 16 positive and 22 negative, after 

 the five-second intervals, and 6 positive and 5 negative, after the ten-second intervals. 



UNCORRECTED VALUES OF LONGITUDE. 



ASSORTED BY LENGTH OP INTERVAL PRECEDING THE SIGNAL. 



Date and signal- 

 station. 



No. 

 cells. 



5^int 

 Pes. 



erval. 

 Neg. 



10' int 

 Pos. 



erval. 

 Neg. 



A 

 5» 



11. 



10» 



'Oct. 25. 



Val. 



Newf. 



10 

 10 



9».075 

 10.235 



9=.060 

 10.265 



9". 033 

 10.308 



8' 975 

 10 332 



9".068 

 10.250 



0».092 

 10.320 



Oct. 28. 



Val. 



Newf. 



10 

 10 



10.406 

 11.678 



10.388 

 11.677 



10.376 

 11.710 



10.445 

 11.742 



10,397 

 11.373 



10.401 

 11.723 



Nov. 5. 



Val. 

 Newf. 



3 

 10 



17.313 

 18.446 



17.287 

 18.467 



17.272 

 18.481 



17.243 



18.482 



17.299 

 18.457 



17.261 

 18.480 



Nov. 6. 



Val. 

 Newf. 



4 

 10 



17.214 

 18.285 



17.220 

 18.298 



17.224 

 18.335 



17.185 

 18.372 



17.217 

 18.292 



17.211 

 18.350 



Nov. 9. 



Val. 

 Newf 



4 

 10 



20.288 

 21.370 



20.281 

 21.350 



20.251 

 21.344 



20.267 

 21.373 



20.284 

 21.359 



20.257 

 21.357 



Hence we may infer the sum of the transmission-times in the two directions to 

 have been 









Excess for 10' interval. | 



Date. 



5» 



10' 



Val. 



Newf. 



Mean. 



Oct. 25 



0=.576 



0'.622 



—0^024 



-fO'.070 



^-0^046 



28 



0.670 



0.716 



—0.004 



+ 0.050 



-1-0.046 



Nov. 5 



0.552 



0.613 



-fO.038 



-1-0.023 



^-o.o6l 



6 



0.469 



0.532 



-1-0.006 



-1-0.058 



-f0.064 



9 



0.469 



0.493 



-1-0.027 



—0.003 



-1-0.024 



Taking next the results afforded by the experiments of November 10 and 16, we 

 find the mean difference between the records of the same signal at the two stations 

 (omitting the 2h 52m as before), to be : — 



