24 THE SCIENTIFIC PAPERS OP 



To describe them here would be a task far exceeding the limits of 

 this paper, and I shall proceed at once to point out what, in my 

 opinion, at least, supported by actual experience, are the best 

 means to be adopted, at the present time, for extending the electric 

 telegraph, both on land and across the seas. 



The foregoing sketch of the gradual development of the electric 

 telegraph, may serve to show that the particular arrangements 

 adopted to indicate or register the message, or the particular com- 

 bination of elementary signs, is of secondary importance, but that 

 every essential progress is marked by the discovery of some new 

 means of generating currents of greater dynamic power, or of 

 producing by their means more decided effects at the further 

 extremity of the conductor. 



Let us inquire, then, what are the conditions of current generator, 

 current conductor and receiver, best calculated to realise a maximum 

 of palpable effect at great distances. 



Inquiry into these questions is of particular interest at the 

 present time, when great efforts are being made to extend tele- 

 graphic communication across the Atlantic and Indian oceans, dis- 

 tances far exceeding the length of any land lines yet constructed. 



Among the different varieties of electricity hitherto applied to 

 telegraphic purposes, that produced by friction possesses the 

 greatest tension or power to overcome resistance in the conductor. 

 But its discharge is instantaneous, and it is, therefore, ill-suited to 

 produce dynamic effects with time or duration for a factor. 



The voltaic current, on the contrary, may be considered as 

 absolutely continuous, and, therefore, as best suited to produce 

 powerful effects, but it is deficient in tension, unless a great 

 number of elements are employed, in which case it becomes expen- 

 sive and troublesome. A battery of sufficient intensity to convey 

 an effect through the Atlantic cable would have to be composed 

 of at least 500 Daniell's cells, according to ordinary practice, but 

 I apprehend that the internal resistance of such a battery would 

 of itself annihilate its presumed power, and that practically no 

 battery of sufficient power could be thus constructed. 



The magneto-electric currents hold an intermediate position 

 between the two just referred to. Their intensity can be increased 

 almost indefinitely, and they are of perceptible duration (the time 

 required to charge an electro-magnet). They may be produced 



