On turning the cylinder C further round, the bit N passes from under the levers, which remain held back by the insertion of the stump in the gatings; and just before reaching the position shown in fig. 59, the slide block R has pushed the bit completely out of the radial slot, and the bit falls down as shown in fig. 59, and drops through a hole into the inside of the safe that is locked. At this point the back pin of the slide block comes in contact with the lower side of the cam groove O in the stump-bolt, fig. 54; and by turning the cylinder C onwards to the position shown in fig. 60, the withdrawal of the bolt B is completed, bringing the parts into the position shown in fig. 60. In these drawings only one lever L is shown; but there are altogether six levers, as shown in the sectional plan, fig. 56. The pin P is fixed in the tail D of the main bolt, so as to travel with the bolt; and by this means the springs I are released from strain, as shown in fig. 60, as soon as the bolt is withdrawn.
From the nicety with which the various parts of this lock are constructed, it is evident that the levers must be very accurately lifted by the bit of the key in order to withdraw the bolt; and therefore any error in the bit, such as would occur with a false bit, will effectually prevent the lock from being opened. This may be illustrated by supposing the false bit to be so close an imitation as to have five of its steps absolutely correct, and the sixth only slightly wrong: though it is almost impossible that such a near approach to correctness could be attained in practice. The counterfeit bit being inserted in the lock, and the cylinder turned round, all will go on the same as with the true bit, up to the time when the false bit reaches the point T of the levers, as previously shown with the true bit in fig. 58. Here a change of action takes place; but what is the nature of the change the operator has no means as yet of ascertaining. In the case supposed, where five of the steps in the bit are right, but the sixth is wrong, the gating of the sixth lever does not precisely coincide with the others, nor with the stump S; and the consequence is that, at the critical moment when the stump ought to spring into the gatings and hold back the levers from falling forwards, it will be prevented from entering the gatings, owing to the entrance being partly blocked up by the one lever, which stands more or less across it.