156 ANNUAL OF SCIENTIFIC DISCOVERY. 



lute time of transmission. This comparison gives us the interval 

 T T" (the difference of the time indicated at the same moment 

 by the two clocks, diminished by the time of transmission in the 

 case of signals given from Valencia, and increased by this 

 amount for signals from Newfoundland). Any excess of the 

 time consumed in the passage of either class of signals should 

 manifest itself by a superior value in the measures of the tem- 

 porary clock-difference derived from that class when the signals 

 are sent westwardly. For earthward signals the reverse holds. 

 An examination of the recorded results shows that the positive 

 and negative signals travel with the same velocity under the 

 same circumstances. 



The speed of the two kinds of signals being thus taken as the 

 same under similar circumstances, the time required for their 

 transmission is easily deduced, being one-half the difference be- 

 tween the measures of longitude as derived from the records at 

 the respective stations. The weak point in the observations is 

 the absence of any automatic record of signals received, but it is 

 probable that the aggregate personal error of the two observers is 

 very close to 0.606% which value is adopted in the investigation. 



The experiments of November 5th and 6th were conducted with- 

 out the use of any earth circuit. Each station sent signals with a 

 battery of 3 MinottPs cells, receiving them with its battery dis- 

 connected. The mean interval consumed in the transmission of 

 the signals appears to have been 0.29 s on the former, and 0.26 s on 

 the latter occasion. 



With a battery of 3 Minotti's cells, the maximum permanent 

 current would not exceed 168 farads in the joined cables, and to 

 develop nine-tenths of this current more than 1^ second would be 

 needed. With 3 DanielPs cells the maximum current would not 

 exceed 185 farads. Assuredly, we cannot suppose that in the 

 lapse of three-tenths of a second, when not more than one-seventh 

 of this current had been developed at the farther station, this bat- 

 tery would have charged the 2 joined cables, each of which pos- 

 sessed an electrostatic capacity of more than 650 farads. Hence 

 the impulse on which the transmission of the signal depends must 

 have been propagated along the conductor by some other means 

 than by charging its successive parts electrically, that is, fully 

 and in the ordinary sense of this expression. The 30 farads, more 

 or less, which could have been generated before the signal arrived 

 at the distant extremity of the cables, would have been consumed 

 in charging the first six or seven hundredths of the conductor. 



Messages were effectually and distinctively transmitted in each 

 direction, by the use of an electromotor formed by a small percus- 

 sion cap containing moistened sand, upon which rested a particle 

 of zinc. The current here evolved could scarcely have amounted 

 to more than 6 or 7 farads, so that nearly 2 minutes would have 

 been requisite for charging a cable, yet the transmission time was 

 certainly very small, although it was not definitely measured. 



The experiments of November 8th and 9th differed from those 

 of the 5th and 6th only in that the Newfoundland battery con- 

 sisted of 10 cells instead of the same number as was employed at 



