PROM THE MAGNETISM OF THEIR BALANCES. 355 



Magnetic Poles exceeds it ; but still the local attraction ope- 

 rates without any increase of force, excepting what may arise 

 from the little augmentation of the magnetic intensity of the 

 earth in those regions ; so that, in reality, the rate of a Chro- 

 nometer in polar regions, where the earth's magnetism acts 

 nearly at right angles to the plane of the balance, could the ef- 

 fect of temperature on the instrument be perfectly compen- 

 sated, ought to be more equable than in any other region, 

 where the direction of terrestrial magnetism is more nearly in 

 the plane in which the balance vibrates. 



In the important and truly scientific experiments of Mr 

 Baklow, on the effects produced in the rates of Chronometers 

 by the proximity of masses of iron, we liave a corroboration 

 of the preceding opinions ; for Mr Barlow, though he obser- 

 ved that a variation of rate was occasioned by the influence of 

 a mass of iron equivalent to the local attraction of a ship, 

 found by no means so great effects as those observed by Mr 

 Fisher. But the force of terrestrial magnetism acting upon a 

 balance that is magnetic, is fully sufficient to account for every 

 change of rate observed. 



Mr S. Varley, in a paper in Tilloch's Philosophical Magazine, 

 published in the year 1798, was the first, I believe, who showed 

 that an irregularity observed in the rate of some time-pieces, was 

 owing to the magnetic state of their balances. He was direct- 

 ed to the inquiry by a watch of excellent workmanship that 

 he had in his possession, which performed the most irregular- 

 ly of any watch he had ever seen. It occurred to him that 

 the cause might be magnetism ; and, on examining the balance, 

 he found it so strongly magnetic, that when suspended hori- 

 zontally without the spring, it directed itself like a compass- 

 needle in a certain position, which it invariably returned to 

 when it was disturbed. The pendulum spring being put on, 



and 



