XVII. On some recent Improvements in Clock-Escapements. By Edmund Beckett 

 Denison, Esq., M.A., of Trinity College. 



[Read Feb. 7, 1853.] 



Foe many years past the great problem of clock-making has been the invention of a simple 

 gravity escapement, which will give a constant impulse to the pendulum, and without any 

 friction capable of sensibly affecting the arc of vibration. For if that can be done, it is 

 evident that all mechanical causes of disturbance of the time of vibration will also be removed ; 

 and if the escapement can also be made so that its effect on the pendulum will still remain the 

 same, notwithstanding any small variation of arc which may arise from changes in the density 

 of the air, or any accidental causes, then the pendulum (so far as it depends on the escape- 

 ment) must remain isochronous as long as it is kept going at all. 



Although Graham's well-known dead escapement, together with that modification of it 

 called the pinwheel escapement, has now held its place against all rivals for more than a 

 century, as the best for astronomical and turret-clocks, it is very far from satisfying the 

 above conditions. Its success is due to a very different cause ; viz. to the fact that the several 

 errors, which arise both from the inconstancy of the impulse and the variations of its large 

 amount of friction, have, by a fortunate coincidence, a tendency to counteract and compensate 

 each other, and sometimes exactly do so. Mr Airy shewed, in a paper on the Disturbances 

 of Pendulums, in Vol. III. of these Transactions, that from this cause the difference between 

 the time of a free pendulum, and of the same pendulum under the influence of this escape- 

 ment, is less than the difference produced by any other escapement. It does not follow indeed, 

 as he thence inferred, that the variation of this difference from day to day (which is the 

 ' rate' of the clock) must be less with this escapement than any other; and in fact (as I 

 shewed in my paper read here in 1848, Vol. VIII. p. 633, and in the Rudimentary Treatise on 

 Clocks afterwards published,) this variation may be made to vanish altogether in what is 

 called a gravity or remontoire escapement, by a construction which makes the difference of 

 time from that of a free pendulum a maximum. But I also shewed that this compensation of 

 errors in the dead escapement may be traced still farther, into the expression for the variation 

 of the difference from a free pendulum ; and that the good going of these clocks depends in a 

 great measure upon two different errors, which arise from the changes of friction in different 

 parts of the clock, also tending to counteract each other*. 



But it must be remembered that all these self-acting corrections are at the best only 

 approximations to accuracy ; especially as the different errors have no constant or definite 



• I take this opportunity of saying that there are several 

 typographical errors in my former paper. The most material 

 one is the omission, in the last page but one, of a* in the 



denominator of the final expression for 86400 d A, or the 

 — daily rate of a gravity escapement. 



