227 



Having, on a former occasion, explained the principle of 

 this instrument, and given the details of its construction, it re- 

 mains only that I should now describe the observations made 

 for the purpose of testing its performance. I shall pass over 

 for the present those which relate to the absolute inclination, 

 because they have yielded results which can be regarded only 

 as approximations to the truth, and I have not succeeded as 

 yet in tracing the errors to their source. It is manifest, how- 

 ever, that an instrument may be a good differential instru- 

 ment, while it is incapable of yielding absolute results ; and 

 there are special reasons why this should be the case with 

 the apparatus now under consideration. Accordingly its 

 failure in the latter respect, even though estabHshed, would 

 furnish no ground for despairing of its success in the former. 



It is obvious that the apparatus is wholly free from the 

 sources of error already noticed, belonging to magnetical in- 

 struments moving on a fixed axle ; and the only doubt of its 

 performance must relate to the changes of induced mag- 

 netism in the iron bar. Thus it might be questioned, before 

 trial, whether such a bar receives in all cases an amount of 

 free magnetism proportional to the inducing force; — whe- 

 ther, again, the minutest changes in the latter are accompa- 

 nied by corresponding changes in the former; — and whether, 

 lastly, the changes thus produced are instantaneous, or, at 

 least, demand no appreciable time for their development. 



In the first experiments which I made, for the purpose of 

 determining these questions, the induced magnetism of the 

 iron bar was altered by means of a permanent magnet, placed 

 in the same right line with the bar, and at a known distance 

 from it. The effect produced upon the position of the sus- 

 pended magnet being observed, the distance was altered by 

 a known amount, and a new observation taken; and so on, 

 at many diflerent distances. Then, the law of action ot the 

 inducing magnet being known, we may calculate the changes 

 of deflection of the suspended magnet, on the supposition 



