176 Performance of Some Timekeepers. 



completely free from the influence of the train. It works 

 without oil, and though delicate in construction, it gives the 

 best results as a timekeeper where the watch is kept in 

 one position, as in marine chronometers. In pocket watches, 

 however, it is, in my opinion, a mistake, and is not so trust- 

 worthy as the lever. 



Although many other escapements have been proposed 

 and constructed — indeed, scarcely any year passes without 

 a new one being invented — yet the five just mentioned are 

 now the only ones in common use. We will therefore pass 

 on to the consideration of the hairspring and balance. 



As mentioned before, Hooke was the first to study the 

 action of springs. He found that the amount of bending 

 of a spring was proportional to the weight that produced it, 

 or, as he expressed it in the learned language of the time, 

 "ut tensio sic vis" From this property, then, when a 

 spring is set vibrating, all the vibrations, whether long or 

 short, ought to be isochronal — that is, performed in the same 

 time. If Hooke had carried out his experiments more 

 carefully, he would have found that in practice this did not 

 strictly apply. The complete investigation was carried out 

 by the celebrated French horologist, Pierre Le Roy, who, in 

 the year 1766, announced his discovery in the following 

 terms : — " There is in every spring, providing it be long 

 enough, a length that causes all the vibrations, whether long 

 or short, to be isochronal. Having fixed upon this length, 

 if you shorten the spring the long vibrations will be quicker 

 than the short. If, on the other hand, you lengthen it, the 

 short arcs will occupy less time than the long." The great 

 object in a good timekeeper, therefore, is to secure a good 

 spring, determine its isochronal point, fix it there, and attach 

 to it a balance of the proper weight and size, so that its 

 inertia, combined with the elasticity of the spring, may 

 cause them to vibrate in the interval of time required by the 

 train, generally one-fifth of a second in a watch and one- 

 fourth of a second in a marine chronometer. No alteration 

 in the length of the spring should be made after this. Any 

 change of rate that may be required should be brought 

 about by changing the weight or dimensions of the balance. 

 In marine and some pocket chronometers this procedure is 

 strictly observed. In watches, however, change of rate 

 is generally made by altering the length of the spring, as 

 any interference with the balance of a watch that has to go 

 in all positions would be likely to produce much larger errors 



