Proceedings of the Royal Society. 26^ 



nute in themselves, appeared, in the autlior's opinion, to point out a 

 species of magnetic action not hitherto described. It has long 

 been well known that striking, twisting, or filing iron, in diflferent 

 directions, with regard to the magnetic axis, materially influences 

 its polarity, but it does not appear to have been remarked that the 

 simple rotation of iron in different directions has any such in- 

 fluence. This, however, the author has ascertained to be the case, 

 and that the laws which govern this peculiar action are so regular, 

 that there can remain no doubt of a corresponding regularity in 

 their causes. 



The attention of the author was first drawn to these phenomena 

 by some apparent anomalies in the magnetic action of an iron 

 plate on the compass, observed in the course of a different in- 

 vestigation. In order to avoid or allow for the disturbing in- 

 fluence of partial magnetism in the iron, it became necessary to 

 attend minutely to the position of certain points in its circum- 

 ference, which corresponded to the maxima and minima of this 

 magnetism. It was then found that these points were not constant, 

 but shifted their position as the plate was made to revolve in its 

 own plane ; or, in other words, that a plate which, in a given 

 position, produced a certain deviation in a compass, no longer 

 produced the same deviation after making an exact revolution in 

 its own plane, although brought to rest, and every part of the 

 apparatus restored precisely to its former place. 



It appeared from this, that the revolution of the plate in its 

 own plane had an influence on its power of deviating the needle 

 independent of the partial magnetism of particular points in it, 

 and the justice of this idea was proved by giving it a rotation in 

 an opposite direction, when the effect on its directive power was 

 ;also reversed. 



The change produced by rotation in the directive power of the 

 plate was found to be a maximum when its plane was parallel to 

 the line of dip on the magnetic axis, and at the same time as little 

 inclined to the horizon as this condition would allow ; but when 

 the plane of the plate was parallel to the horizon the effect was 

 .diminished in the ratio of 5 to 1 } and when perpendicular to the 



