229 



The piece of railing which was the subject of experiment 

 was raised, by the help of a windlass and pulleys, into the 

 horizontal plane containing the magnet, and was fixed in the 

 required position at the distance of 60 feet. The position of 

 the magnet being observed, the mass of iron was lowered, and 

 removed to a sufficient distance upon a truck, and a fresh ob- 

 servation taken. This process was repeated several times in 

 rapid succession, and a series of observations thus taken, with 

 the magnet alternately disturbed and undisturbed. The fol- 

 lowing are the results : 



These observations are very satisfactory. They give, 

 for the mean difference of the readings due to the presence of 

 the railing, + 1-96 divisions of the scale = + l''4\. Hence the 

 length of the piece of railing being 14 feet, the effect pro- 

 duced by a piece whose length = 1 foot, is 



£ = + 0'-101. 



But D = 60, and a = 153, the unit of length being I foot. 



Also a = 31° ; and the length of the line of railing being 255 



feet, /3 = tan-i (1-67) = 59°. Hence the quantity within the 



brackets, in the value of u, is equal to - 0-363 ; whence 



finally, 



u = - 0-164. 



The only other disturbing causes, in addition to those 

 already noticed, are those which affect the position of the read- 



