Another receiving instrument employed on most of the longer cables is the siphon recorder of Lord Kelvin, shown in figure 54, which marks or writes the message on a slip of travelling paper. Essentially it is the inverse of the mirror instrument, and consists of a light coil of wire S suspended in the field between the poles of a strong magnet M. The coil is attached to a fine siphon (T5) filled with ink, and sometimes kept in vibration by an induction coil so as to shake the ink in fine drops upon a slip of moving paper. The coil is connected between the cable and the earth, and, as the signal current passes through, it swings to one side or the other, pulling the siphon with it. The ink, therefore, marks a wavy line on the paper, which is in fact a delineation of the rise and fall of the signal current and a record of the message. The dots in this case are represented by the waves above, and the "dashes" by the waves below the middle line, as may be seen in the following alphabet, which is a copy of one actually written by the recorder on a long submarine cable. Owing to induction, the speed of signalling on long cables is much slower than on land lines of the same length, and only reaches from 25 to 45 words a minute on the Atlantic cables, or 30 to 50 words with an automatic sending-key; but this rate is practically doubled by employing the Muirhead duplex system of sending two messages, one from each end, at the same time. The relation of the telegraph to the telephone is analogous to that of the lower animals and man. In a telegraph circuit, with its clicking key at one end and its chattering sounder at the other, we have, in fact, an apish forerunner of the exquisite telephone, with its mysterious microphone and oracular plate. Nevertheless, the telephone descended from the telegraph in a very indirect manner, if at all, and certainly not through the sounder. The first practical suggestion of an electric telephone was made by M. Charles Bourseul, a French telegraphist, in 1854, but to all appearance nothing came of it. In 1860, however, Philipp Reis, a German schoolmaster, constructed a rudimentary telephone, by which music and a few spoken words were sent. Finally, in 1876, Mr. Alexander Graham Bell, a Scotchman, residing in Canada, and subsequently in the United States, exhibited a capable speaking telephone of his invention at the Centennial Exhibition, Philadelphia. Figure 56 represents an outside view and section of the Bell telephone as it is now made, where M is a bar magnet having a small bobbin or coil of fine insulated wire C girdling one pole. In front of this coil there is a circular plate of soft iron capable of vibrating like a diaphragm or the drum of the ear. A cover shaped like a mouthpiece O fixes the diaphragm all round, and the wires W W serve to connect the coil in the circuit.