325 



the same dip, within five centesimal minutes, 

 as immediately after the earthquake. I did 

 not indeed change the poles of the magnet, at 

 each observation, as I have done since in a 

 long series of dips determined conjointly with 

 Mr. Gay-Lussac, in 1805, and 1806, in France, 

 Italy, Switzerland, and Germany ; and as the 

 astronomers did constantly in Captain Cook's 

 second Voyage. This operation is long and de- 

 licate, when we are almost always obliged to 

 make our observations in the open air. On 

 leaving Europe, the Chevalier Borda advised 

 me to deprive the needle of it's magnetic power 

 only after certain intervals, and to take note 

 of the differences. These differences did not 

 amount, in the experiments made with Mr. Le- 

 noir at Paris, to more than 12 minutes ; and at 

 Mexico, in different trials, to 8, 15, 6, and 10 

 minutes: the needle too, which was of well 

 tempered steel, retained it's full polish during 

 five years. Besides, in the phenomenon of which 

 we are treating, the question relates only to a 

 change of apparent dip, not to an absolute quan- 

 tity. Not having touched the needle, I do not 

 see the probability of an error of a centesimal 

 degree. 



It is well known that the shock, in modifying 

 the position of the particles of iron, cobalt, or 

 nickel, modifies also their magnetic properties ; 

 that it is capable of giving polarity, and some- 



