( 363 ) 



APPENDIX. 



OiNCE forwarding my communication to the Royal Society, 

 " On the Errors in the Sea-Rates of Chronometers, arising 

 " from the Magnetism of their Balances," &c. I have con- 

 structed a temporary apparatus on the principle described in 

 the third " suggestion for removing this source of error." I 

 was at first doubtful whether a plate, however light, when 

 loaded with the weight of a pocket chronometer, could be 

 made to traverse by the polarity only of a compass-needle ; 

 and vsrhether, within a moderate compass, the magnet intended 

 for directing the plate could be so far removed from the chro- 

 nometer as to prevent all fear of additional mischief from its 

 proximity. 



On trial, I vras happy to find my apprehensions respecting 

 the interference of both these apparent difficulties entirely re- 

 moved. For, by means of a compass-needle, indifferently 

 magnetised, the plate for the chronometer traversed, when 

 loaded w^ith a pound weight avoirdupois ; and the magnetic in- 

 fluence of this needle, at five inches distance (the distance be- 

 tween the chronometer plate and the needle) was only equal 

 to the directive force of the earth on a horizontal needle in 

 Britain. Now, such a degree of influence would probably be 

 an advantage to the chronometer's going rather than other- 

 wise ; because the denomination of magnetism in either end 

 of the needle, operating on the part of the chronometer to 

 which it was most contiguous, would be of the opposite kind 

 to that of the earth operating on the same part of the chrono- 

 meter ; hence the tendency of the magnetism of the needle on 



the 



