in different Latitudes. 301 



the meridian of Paris; which gives 16-326° (14° 52' 25") 

 ot'latiludc counted from the magnetic equator, and b'S'TSQO" 

 (48" 21' 53") of west longitude proceeding from the node 

 formed by that equator v^ith the equator of t'le earth. The 

 inchnation of the magnetic needle was observed in that 

 place by M. Humboldt in the month of Mcssidor, year 8, 

 and found to be equal to 33'78° of the centigrade division*. 

 A comparison of this result, with the other observations 

 of M. Humboldt, shows that it may be indeed considered 

 as agreeing to that latitude. 



To make use of it I have successively given to K different 

 values in the formula: I have calculated the inclinations 

 resulting from that latitude ; and, comparing these results 

 With that which M. Humboldt really observed, the progress 

 of the errors naturally led me to the most proper supposi- 

 tion. The followino; is a table of these trials : 



The first value of K would place the centre of the matr* 

 netic forces at the surface of the earth and the poles of the 

 magnetic equator. It is seen that this supposition cannot 

 be admitted, because it would give an increase of inclina- 

 tion much less rapid than that indicated by observations. 

 The case is the same with the following results, which 

 place the centres of action on the terrestrial radius at 

 different distances from the centre of the earth ; but it is 

 seen also, in general, that they approach more and more to 

 the truth in proportion as this distance becomes less; which 

 evidently shows that the two centres of action of the mag- 

 netic forces are situated near the centre of the earth. All 

 the other observations of M. ilumboklt would also lead to 

 the same consequence. 



The most proper supposition would be to make K null, 

 or so small that it would be needless to pay attention to it ; 

 which amounts to the same thing as to consider the two 

 centres of action placed, as we may say, in the same mo- 

 lecula. The result, indeed, obtained in this maimer is the 

 most exact of all ; it is equal to 3 i-0643" : this value is still 



* AU tlie measures of inclination which I have given in this me:noir will 

 he er.prc3ii.U, like those of M. HumbuUt, in dcciniiil parts of a qnadriuu. 



a little 



