Performance of Some Timekeepers. 179 



interfered with the isochronal quality of the spring, it was 

 soon abandoned in favour of the compensation balance 

 invented by Julien Le Roy, and afterwards much improved 

 by Arnold and Earnshaw, who left it in very nearly the form 

 it has retained ever since. According to this construction, 

 the balance, instead of being a complete circle, consists of a 

 steel cross bar, to each end of which is fixed a nearly semi- 

 circular strip. These strips are composed of two metals, 

 brass on the outside and steel on the inside. The old 

 makers used to connect these metals by a number of very 

 fine rivets. The modern practice is to melt one on the 

 other. Screws called timing screws are fixed at the end of 

 the cross-bar ; any motion of these simply alters the inertia 

 of the balance, without interfering with the compensation. 

 They are therefore used for changing the rate. Heavy 

 screws or weights are also attached to the bi-metallic strips. 

 These are moved further from or nearer to the cross-bar, as 

 we want to increase or diminish the effect of change of 

 temperature. This balance acts in the following manner : — 

 An increase of heat will cause the brass on the outside to 

 expand more than the steel on the inside. This will cause 

 the strips to curve inwards, and carrying the weights with 

 them they thus diminish the inertia of the balance and 

 compensate for the diminished elasticity of the spring ; a 

 decrease of heat will in a similar manner make them open 

 outwards, and increase the inertia of the balance. This 

 is still the style of balance most generally used. It is called 

 simply a compensation balance. Its action, however, is not 

 perfect. This was first explained by Mr. Dent, in 1833, 

 though its defect had been a matter of observation in good 

 chronometers for some time before. It was found that while 

 the chronometer would go very well for a moderate range of 

 temperature, yet it always lost above or below these limits 

 when the range was considerable ; or, if it went right at two 

 extreme temperatures, it always gained at the mean one. 

 The explanation is this : The effects of temperature on the 

 spring would cause changes in its strength proportional to 

 the changes of temperature. To compensate for this the 

 moments of inertia of the balance ought to change in the 

 same proportion. Instead of this, however, equal variations 

 of heat will cause the compensating weights to alter their 

 distances from the centre of the balance by equal variations 

 of space; but the moments of inertia of the balance are 

 proportional, not to the distances of the weights from the 



o 2 



