So current from a battery flows to the two wires leading to the distant station, each passing through the rheostat connected to one of the cranks. We may think of the rheostats as taps which can be turned on or off by the action of the cranks. Let us imagine that crank _a_ is in the position when the current flows freely--when the electrical "tap" is fully open; then a strong current will flow along wire _a_, returning to the sending battery via the earth. As that crank is moved the current will gradually be reduced, until, if it be moved right over to the other extreme, the current will be at its feeblest. [Illustration: FIG. 15.--A Message received by Telewriter.] Arrived at the other end, this current passes to a device which we may describe simply as a magnet so arranged that its action pulls round a crank against the restraining action of a spring. Now the stronger the current the more does that magnet pull and the farther does the receiving crank turn. The sending crank varies the resistance, the resistance varies the current, the current varies the strength of the receiving magnet, and the magnet varies the position of the receiving crank. Properly adjusted, then, the motion of the crank at the one end is communicated through that long chain of causes and effects, until at last it is repeated _exactly_ by the movement of the crank at the other end. The same thing occurs simultaneously over each of the two wires, crank _a_ at the sending end communicating over wire _a_ to crank _a_ at the other end, while crank _b_ communicates its motion over wire _b_ to the other crank _b_. Each sending crank is closely imitated in its every action by the corresponding one at the distant station. The two receiving cranks are connected by light rods to the receiving pen in precisely the same way that the sending pen is connected. Consequently, not only are the separate movements of the two cranks repeated at the remote station but the complex movements of the sending pen, which gave rise to the actions of the cranks, are also conveyed to, and repeated by, the recording pen. The movements of the first pen are resolved into rotating motions by the two cranks, these are transferred to the other cranks, and their movements are in turn converted back into the written curves. Thus as the pen in the artist's hand draws his sketch, so does the automatic hand at the other place, it may be at a great distance, repeat faithfully his work, and the sketch grows line by line simultaneously at both ends. There is not space here to detail how, by another current superposed upon those referred to already, the receiving-pen is made to dip itself periodically into the inkwell at the will of the sender. By a cunning use of alternating current this is done without in any way interfering with the action of the cranks as described above.