When the pendulum comes back towards the end of its beat, the arm _A H_ strikes the arm _E F_, and raises the ratchet _C D_. This releases the wheel, which has a weight wound up upon it, and therefore at once tries to go round. The consequence is, that the pin _G_ strikes upon the arm _A H_, and thus on its return stroke gives an impetus to the pendulum. As the pin _G_ moves forward it slides on the arm _A H_ till it slips over the point _H_. The wheel now being free, would fly round were it not that when the pendulum returned, and the arm _A H_ was lowered, the ratchet had got into position again and its point _D_ was ready to meet and stop the next pin that was coming on against it. At each blow of the pins against the pendulum a “tick” is made, at each blow of a pin against the ratchet a “tock” is sounded, so that as it moves the pendulum makes the “tick-tock” sound with which we are all familiar. Hence then a clock consists of a wheel, or train of wheels, urged by a weight or spring, which strives continually to spin round, but its rotation is controlled by an escapement and pendulum, so contrived as only to allow it to go a step forward at regular equal intervals of time. But this would make only a poor sort of escapement. For the mode of driving the pendulum adds a complication to the swing of the pendulum. Instead of the pendulum being simply under the accelerative force of gravity, it is also subjected to the acceleration of the pin _G_. This acceleration is not of the “harmonic” order. Hence so far as it goes it does not tend to assist in giving a harmonic motion to the pendulum, but, on the contrary, disturbs that harmonic motion. Besides this, the impulse of the pin is in practice not always uniform. For if the wheel is at the end of a train of wheels driven by a weight, though the force acting on it is constant, yet, as that force is transmitted through a train of wheels, it is much affected by the friction of the oil. And on colder days the oil becomes more coagulated, and offers greater resistance. Moreover, as will be explained more in detail afterwards, the fact that the impulse is administered by _G_ at the end of the stroke of the pendulum is disadvantageous, as it interferes with the free play of the pendulum. From all these causes the above escapement is imperfect in character, and would not do where precision was required. [Illustration: FIG. 41.]