the Magnetism in Steel Bars. 249 



observation to another, and are besides but small, it is extremely 

 doubtful whether, notwithstanding the care taken in adjusting 

 the bars for examination and replacing them afterwards, they 

 were not due to accidental derangement, when the bars were 

 moved at the several times of observation, rather than to a 

 tendency in the magnetism to resume a state of symmetrical 

 distribution. Indeed it is possible that the needle made use of 

 to determine the positions of the poles and magnetic centre 

 may have slightly modified the magnetic state of the bar, at 

 each observation, since it was necessary, for this purpose, to 

 bring it within a very small distance ; and such influence it is 

 scarcely possible to prevent, especially in the case of rather 

 soft bars, as these purposely were. 



The changes in the bars I., III., IV., are in general much 

 greater from July, 1828, to July, 1829, than subsequently, and 

 are the most indicative of the mutual action of the poles upon 

 each other, since the changes in the positions of the more dif- 

 fused poles are much greater than could have arisen from 

 errors of adjustment or observation ; but even these, admitting 

 them to be entirely due to such action, are so small, consider- 

 ing the time during which they took place, that the force which 

 produced them must have been almost evanescent as compared 

 with the coercive force of the steel. 



In the bar II., the changes are extremely small from July, 

 1828,. to July, 1829, and also from the latter date to June, 

 1830, and, varying much in their direction, do not indicate 

 that they were caused by internal action ; but from June to 

 December, 1830, a decided change in the positions of the poles 

 and magnetic centre took place. This, however, I have no 

 doubt was entirely due to accident. When I removed this 

 bar in December, for the purpose of making observations on it, 

 in order to see which was the marked end, I incautiously ap- 

 proached a candle unfortunately placed upon a box containing 

 a very powerful magnet, and, from the small amount of all the 

 other changes, I have no hesitation in attributing the consider- 

 able change here observable to the action of this large magnet 

 upon the poles of the bar. Those who have not been exten- 

 sively engaged in delicate magnetical experiments can scarcely 

 be aware of the difficulty of guarding against such accidental 



