And thus we see the whole scheme. We regulate the speed by the rheostat, and meanwhile that tell-tale stream of alternating current comes flowing out of the motor to indicate to us what the speed is, while the "frequency meter," with its various vibrating bars, interprets to us the message which the alternating current brings to us. So by watching the meter we know when we have got the speed that we desire. But even that is only half the battle. We have seen how to make a machine turn at any desired speed, and so we can adjust any two, so that they revolve at the same speed, but we have not seen how to start and stop the two machines at the same time. First of all, it must be understood that in the case of the receiving machine there is a friction clutch, as it is termed, between the motor and the cylinder which it is driving. That means that while, under ordinary circumstances, the motor drives the cylinder round, we can, if we like, hold the latter still without stopping the motor. When we do so, the connection between the two simply slips. So if we fit a catch on the cylinder which is capable of holding it from rotating, we can still start the motor, and the latter will work. Then, the moment the catch is released the cylinder will begin to turn too. The commonest form of "friction drive" is the flat leather belt upon two pulleys, which everyone has seen at some time or other in a factory. And it will be quite easy to conceive how, if one of the driven machines were to stick, the belt might simply slip upon one of the pulleys, yet, as soon as the machine became free again, it would rotate just as it did before. It is just the same with what we are considering. The motor works continuously at its proper speed, but the cylinder can be stopped when desired by the catch. Combined with the catch is an electro-magnet, and through its coils there flows the current of electricity which is engaged in printing the picture on the cylinder. If a magnet be arranged to attract another magnet, it will do so only when the energising current flows one way. When it flows the other way, it does not attract. Therefore it is easy to arrange matters so that the printing current, though passing through the coil of the magnet, shall not pull open the catch. But if that current be _reversed_ in direction for a moment the magnet gives a pull, open flies the catch, and away goes the cylinder upon its revolution. Thus, we see, all that is necessary to start the receiving cylinder is to reverse the current for a moment.