1820.] Mr. Barlow's Essay on Magnetic Attractions. 306 



pole formed at some distance from the contact, and beyond this a faint south pole. 

 These gradually advance along the bar. The remote extremity becomes first n 

 south pole, and it is not till after a very long while (if ever) that it becomes a sim- 

 ple vigorous north pole: more commonly it remains a diffused and feeble north 

 pole : nay, if llie bar be very long, it often happens that we have a succession of 

 north and south poles, which never make their way to the far end of the bar." 



Now let me ask again, if there be (and no one who lias made experiments will 

 doubt the fact) the difficulty above slated, in transmitting the magnetic quality 

 through a bar of hard steel, even when in contact with a powerful miignet, how 

 are we to account for that sudden change which (according to the explanation 

 given of the former experiment) lakes jilace merely from a change of position ? At 

 all events I tliink it must be admitted that the direction given for using soft iron is 

 unnecessary, and worse than superfluous, because it supposes a distinction which 

 does not exist. 



143. I have also some doubt whether the computation given by M. Biol relative 

 to the position of the poles of the terrestrial magnet does not go a considerable way 

 towards proving a defect in the theory in question. 



If the earth be a magnet, and has, like other magnels, two sirongly allracting 

 poles, we ought to find a very considerable increase of magnetic intensity, as we 

 approach towards those parts of the globe where they are situated : whereas, as far 

 as voyages have hitherto been made, the change in this respect is far from being 

 inconsiderable. 



This circumstance alone would naturally lead us to infer, that they must be 

 situated atan immense depth below the surface; but we have better authority with 

 reference to this point. JM. Biot, after assuming their existence, left the distance 

 between them indeterminate ; nnd then, by comparing his general expressions with 

 the results of observations in different latitudes, endeavoured to determine their 

 distance and position. 



Making this comparison he found that the nearer tiiese poles were supposed to 

 approach towards each other, the nearer his numbers approximated to those de- 

 duced from observation ; and, finally, by assuming the poles infinitely near to each 

 other in the centre of the earth, he obtained something like coincidence between his 

 two series of results. 



Now 1 cannot help observing, that it appears to me if this computation prove any 

 thing, it is, that the hypothesis of the earth, containing within itself two magnetic 

 poles, is altogether erroneous ; for what idea can we have of an infinitely small 

 magnet, having its north and south poles coincident in the centre of the earth, giv. 

 ing directions to bodies at the distance of 4000 miles ? 



Whether there are any circumstances connected with the experiment reported in 

 Art. 141, or in the nature of the deduction referred to above, that have escaped my 

 observation, and that will reconcile tliese apparent anomalies with the theory in 

 question, 1 am unable to say ; but till these circumstances are pointed out and ren- 

 dered obvious, I cannot but entertain some doubt with respect to the accuracy of 

 the principles upon which that theory is founded. 



144. 1 have said that M. Biol, by assuming the two poles of the terrestrial mag- 

 net to be indefinitely near to each other in the centre of the earth, had found 

 " something like coincidence between his computed and observed result." This 

 was meiint principally in reference to the law of the dip, which, according to the 

 corrected, or rather simplified formula of that author, is 



tan. 8=2 Ian. X, 



where 8 is the dip and K the magnetic latitude. 



145. In order to enable the reader to judge of the approximation which (his law 

 affords towards the angle of inclination, as determined by the best observalions, 

 Ihave made the computations indicated in the following table: 



Vol. XVI. N° IV. U 



