THE TRANSATLANTIC LONGITUDE. 



91 



circuit signals, lead to the singular inference that the latter travelled most rapidly 

 in the case of Newfoundland signals with the Valencia battery, while for the 

 Valencia signals with the Newfoundland battery precisely the reverse was the case. 

 For this I have no explanation to suggest. It has been impossible for me to shake 

 off a suspicion that the same error in the connections on November 10, which 

 occasioned the discordances heretofore mentioned, may have also acted to produce 

 the discrepancies here manifested ; but I will confine myself to a statement of the 

 results, and leave any possible reconciliation of discrepancies for the future. 



There are two ways in which the comparative velocity of these two sorts of 

 signals may be examined. One is by comparing the values of the approximate 

 longitude, as given by the make-circuits and break-circuits respectively, for which 

 purpose all corrections for clock-error, &c. may be disregarded, since they affect 

 both sets of signals alike. The other is by deducing the sum of the transmission- 

 times for each kind of signals taken together with the signals sent in the opposite 

 direction. This latter method permits the employment of a mucTi larger number of 

 observations, and by use of the value of the transmission-time for the positive and 

 negative signals, as previously deduced, it allows a tolerably approximate determi- 

 nation of the actual time for the signals in question. The former gives only the 

 difference between the intervals consumed by the two classes respectively, but it 

 affords measures of this difference free from the influence of extraneous sources of 

 error. I will state the results obtained by each of these methods. 



Beginning with the first named, it will readily be perceived that an excess in 

 the approximate longitude, as deduced from make-circuit signals indicates an infe- 

 rior velocity for these, when they are sent from Newfoundland eastward, but a 

 superior velocity when they are sent from Valencia westward. Yet such an excess 

 is manifested in both cases, as will be seen from the appended table. 



Exp't. 



Se 

 Date. 



RIES I. — S 



Earth-con- 

 nection. 



SIGNALS 



Mak 

 No. 



PROM NeWFOTJNDL 



e-cirouit signals. 

 Mean interval. 



and; B 



Brea 

 No. 



ATTERT AT VaLEN 



t-circuit signals. 

 Mean interval. 



CIA. 



Excess for 

 make- 

 circuits. 



No. of 

 cells. 



I. 8 

 I. 1 

 I. 2 

 I. 3 



Nov. 1 

 10 

 10 

 10 



None 

 Middle 

 Zinc 

 None 



2 



3 



9 



10 



2" 52" 15^515 

 22,293 

 22.293 

 22.091 



4 



4 



10 



10 



2'' 52" 15'. 335 

 21.997 

 22.040 

 21.965 



+ 0M80 

 + 0.296 

 + 0.253 

 + 0,126 



20 

 4 

 4 

 4 



Series III. — Signals from Valencia; Battery at Newfoundland. 



III. 1 



Nov. 16 



Middle 



9 



III. 2 



16 



Zinc 



10 



in. 3 



16 



None 



10 



20».950 

 21.002 

 21.082 



20^.798 

 20.880 

 21.030 



— 0M52 

 —0.122 

 —0.152 



The results by the second method of inquiry may be obtained by assuming the 

 transmission-time for signals from Valencia, November 1, to have been 0'.214, and 

 that for signals from Valencia, November 10, and from Newfoundland, November 16, 

 to have been 0'.264 ; Thus we have : — 



