In telegraphy messages not only may be received, but also recorded, by the Morse printer or one of its modifications, but in ordinary telephony there is no mechanical method of recording messages. This means that we can communicate by telephone only when we can call up somebody to receive the message at the other end, and if no one happens to be within hearing of the telephone bell we are quite helpless. This is always annoying, and if the message is urgent the delay may be serious. Several arrangements for overcoming this difficulty by means of automatic recording mechanism have been invented, but the only really successful one is the telegraphone. This instrument is the invention of Waldemar Poulsen, whose apparatus for wireless telegraphy we shall speak of in the next chapter. The telegraphone performs at the same time the work of a telephone and of a phonograph. In the ordinary type of phonograph the record is made in the form of depressions or indentations on the surface of a cylinder of wax; these indentations being produced by a stylus actuated by vibrations set up in a diaphragm by the act of speaking. In the telegraphone the same result is obtained entirely by electro-magnetic action. The wax cylinder of the phonograph is replaced by a steel wire or ribbon, and the recording stylus by an electro-magnet. The steel ribbon is arranged to travel along over two cylinders or reels kept in constant rotation, and a small electro-magnet is fixed midway between the cylinders so that the ribbon passes close above it. This magnet is connected to the telephone line, so that its magnetism fluctuates in accordance with the variations in the current in the line. We have seen that steel retains magnetism imparted to it. In passing over the electro-magnet the steel ribbon is magnetized in constantly varying degrees, corresponding exactly with the variations in the line current set up by the speaker’s voice, and these magnetic impressions are retained by the ribbon. When the speaker has finished, the telephone line is disconnected, the ribbon is carried back to the point at which it started, and the apparatus is connected to the telephone receiver. The ribbon now moves forward again, and this time it acts like the speaker’s voice, the varying intensity of its magnetic record producing corresponding variations in the strength of the magnet, so causing the receiver diaphragm to reproduce the sounds in the ordinary way.