452 Mr. A. T. Hare [Feb. 21, 



stop which limits the travel of the gravity piece. The latter, there- 

 fore, goes on impelling the pendulum until it is brought up against 

 the armature. When this happens the gravity piece is left behind 

 by the penduhim and the circuit is broken. At once the armature 

 falls against a stop, and the gravity piece is Hfted, so that the 

 pendulum takes it up again at a higher level than that at which they 

 parted company. Sir David Gill found trouble from the slight 

 adhesion which exists between two metallic surfaces when a current 

 is broken between them, and gave much attention to experiments 

 designed to avoid this. I do not know how far he succeeded, but it 

 seems clear from Prof. Sampson's paper that the escapement is very 

 successful now. The idea has probably occurred to many people. I 

 began making a clock about thirty years ago on what w^as practically 

 the same principle, but gave it up because at that time it did not 

 seem practicable to find a battery capable of giving a current lasting 

 nearly half a second for each second that passes. 



Another of Prof. Sampson's clocks is driven by an escapement 

 invented by Riefier, of Munich, which is unlike any of those we have 

 been considering, and in which the necessary energy is communicated 

 to the pendulum by bending the suspension spring. The block from 

 which the suspension spring hangs, instead of being fixed as immov- 

 ably as possible, which it generally is, is supported on knife-edges, 

 and the suspension spring, which, of course, always tries to keep 

 straight, causes the block to turn on these edges, and so unlock the 

 scape-wheel, whicli bends the spring back against the motion of the 

 pendulum and thus keeps it going. 



The third escapement, which is being observed at Edinburgh, and 

 the last I propose to refer to, is that adopted by the Synchronome 

 Co., and belongs to the class where the action takes place at the 

 bottom of the pendulum or of the crutch instead of the top. This is 

 fully described in the specification of a patent granted to Mr. Shortt, 

 and numbered 9527 of 1915. 



So much for escapements. 



We may, in conclusion, for a moment review the difficulties 

 attending the accurate measurement of time and note how they have 

 been attacked. 



If ever a perfect clock is constructed it will certainly be a 

 pendulum clock, and it will have to fulfil two conditions, necessary 

 and sufficient. They are these : — First, the moment of inertia of the 

 pendulum must be invariable ; and, secondly, the forces which act on 

 it must be invariable. If these two conditions could be fulfilled, the 

 last word in horology would have been said. So far, of course, 

 neither condition has been fulfilled, but surprisingly good work has 

 been done. As for the first condition, that the moment of inertia 

 uuist be invariable, the chief difficulty is to avoid change by change 

 of temperature. There are two ways of diminishing this change. 

 The pendulum must be compensated in one of the well-known ways 



