324 Rev. Mr Scoresby on the Uniform PermeaUlity of 



When the action of the magnet on the compass is found to be 

 very small, or when particular accuracy is required, the sum of 

 the deviations, produced by the two poles of the magnet M in 

 succession, is taken, instead of the single deviation, by which 

 not only is the angle doubled, but the error of observation re- 

 duced to one-half. 



The first verification experiments, for ascertaining the degree 

 of accuracy with which distances might thus be determined, 

 using only ordinary instruments, namely, a pocket-compass 

 and a twelve-inch bar magnet, were made through the walls 

 and furniture, &c. of my own house. This magnet, acting on 

 the compass through a partition wall of brick and plastering, 

 produced a deviation in the compass of 45°. The comparative 

 experiment, when the magnet was brought round into the room, 

 and placed in the opposite position as to the compass, gave a 

 similar deviation at the distance of 19i inches : the distance 

 through the wall was then ascertained to be 18i| inches, making 

 an error of |th of an inch, or ^^th of the whole distance. A similar 

 experiment was then made through a brick wall and book- case, 

 filled with books, when the comparative experiment gave the 

 thickness 2 1 1 inches, whilst the real thickness was found to be 

 21 1 ; errors ^th of an inch, being only ^j^th part of the whole 

 distance. These results encouraging me to proceed, I subse- 

 quently made the experiments contained in the following table, 

 using only the pocket-compass for the deviations, and two pairs 

 of bar magnets, one pair (B), of two feet in length, and the 

 other (C), of 12 inches, for giving the magnetic influence. 



