intensity due to the arrival of a current and the gradual decrease due
to the ceasing of that current. The possible speed of signalling was
found to be very nearly proportional to the squares of the lengths
spoken through. Thus, a speed which gave fifteen dots per minute in a
length of 2191 nautical miles reproduced all the effects given by a
speed of thirty dots in a length of 1500. At these speeds, with ordinary
Morse signals, speaking would be barely possible. In the Red Sea, a
speed of from seven to eight words per minute was attained in a length
of 750 nautical miles. Mechanical senders, and attention to the
proportion of the various contacts, would materially increase the speed
at which signals of any kind could be transmitted. The best trained hand
cannot equal the accuracy of mechanism, and the slightest irregularity
causes the current to rise or fall quite beyond the limits required for
distinct signals. No important difference was observed between signals
sent by alternate reverse currents and those sent by the more usual
method. The amount of oscillation, and the consequent distinctness of
signalling, were nearly the same in the two cases. An advantage in the
first signals sent is, however, obtained by the use of Messrs. Sieman's
and Halske's submarine key, by which the cable is put to earth
immediately on signalling being interrupted, and the wire thus kept at
a potential half-way between the potentials of the poles of two
counter-acting batteries employed, and the first signals become legible,
which, with the ordinary key, would be employed in charging the wire.

A system of arbitrary characters, similar to those used upon the Morse
telegraph, was employed, and the letter to be indicated was determined
by the number of oscillations of the needle, as well as by the length of
time during which the needle remained in one place. The operator, who
watched the reflection of the deflected needle in the mirror, had a key,
communicating with a local instrument in the office, in his hand, which