88 THE TRANSATLANTIC LONGITUDE. 



four days are concerned, we might infer that on the complete metallic circuit 

 formed by the two cables, the time for transmitting the signals through about 3476 

 kilometers, or 2160 statute miles, was not far from 0\29 for a battery of 3 cells, 

 0'.26 for one of 4 cells, and 0\215 for one of 10 cells. 



On the other hand, the average transmission-time for signals sent by a current 

 induced in a single cable, by means of a " condenser" with a battery of 10 cells, 

 was 0\31 on the 25th, and 0\34 on the 28th October; the mean interval for 

 these two days being 0".328. Each of the condensers used possessed an electro- 

 static capacity of about 20 farads; so that with a tension of 10 cells, or 8.4 volts, 

 their capacity would be not far from 168 farads, or equal to that of about 590 

 miles [945 kilometers] of cable — in other words, a little more than one-quarter 

 of the capacity of one Avhole cable. 



The value of those experiments in which the batteries were connected with the 

 earth is seriously impaired by the series of mistakes made at Newfoundland on the 

 10th November. On that day 20 cells were used instead of 4, and the prescribed 

 connection of the batteiw with the ground was forgotten, so that both the electro- 

 motive and the electrostatic relations became too complicated for any safe inferences 

 as to the results. But apart from these, some other grave error appears to have been 

 committed, by which we are apparently led to the singular result that the average 

 time consumed in the transmission of signals was 0'.31 for the positive, and only 

 0'.24 for the negative signals ; although the only difference between these classes 

 consisted in an interchange of electrodes relatively to the two cables, and although 

 the transmission-time for the two cables is shown by all our other experiments to 

 be practically equal. The sole reason which I can discover for any difference be- 

 tween these two kinds of signals seems inadequate to explain the phenomenon, yet 

 it ought not to be overlooked. It is this : — 



The construction of the signal-keys was such that, in the only manner in which 

 it was safe to use them for these experiments, the battery-circuits remained con- 

 nected Avith the cables at the receiving station. The cables were connected with 

 each other without the battery, and the battery was short-circuited independently 

 of them ; still, a metallic connection did exist between the telegraphic circuit which 

 was formed by the two cables together with their transatlantic battery, on the one 

 hand, and the temporarily disused (and also closed) local circuit, on the other. So 

 long as there is no earth-connection in this local circuit, its effect may fairly be left 

 out of all consideration ; but whenever any such connection is introduced, the case 

 is changed. 



In the experiments of Nov. 10, the zinc of the 4-cell battery at Valencia was 

 provided with an earth-connection, while the 20-cell battery at Newfoundland was 

 insulated. And, since the galvanometer at each terminus was situateci upon that 

 cable to which the platinode was applied for those signals which Ave term positive, 

 some difference must have existed in the action of the two classes of signals from 

 NeAvfoundland upon the Valencia galvanometer. For the NeAvfoundland signals 

 Avould exert a tension on the cable of 1866, Avhich on reaching Valencia might act 

 for an instant inductiA^ely upon the local circuit, before the dynamic equilibrium of 

 the main circuit should be established by means of the opposite tension upon the 



