98 On the LavDs of Ttrreslrial Magnetism 



this magnet, so as to cause the anomalies which occur on thk 

 side of the globe to disappear, and also to make the small dips 

 that have been observed to take place in the southern part of 

 the South Sea, agree with the large d\]n which take place in 

 North America. By thus dividing some other secondary centres 

 in the points of the globe where the irreguktrities of the variations 

 appear most extraordinarv, it is probable that we should end, by 

 representing all of them with exactness, as also the dips and in- 

 tensities. Thus it is in the system of the world : the principaJ 

 motion produced by the action of the sun, is modified by the 

 perturbations that are caused by the small masses of the planets. 

 But as it is necessary to know the places of these masses, in or- 

 der to calculate their influence, so likewise it is necessary that 

 the most accurate ol)servations should point out the position of 

 the different secondary magnetic centres, before we can compute 

 their effects. 



Is the central magnetic action, indicated by these phajnomena 

 with so much probability, really produced by a magnetic nucleus 

 contained in the interior of the terrestrial globe ; or, is it not 

 the principal resultant of all the magnetic particles disseminated 

 in its substance ? Of this we are ignorant: but the last sup « 

 position appears to be the most probable. The secondary cen- 

 tres would then be determined by some local attractions be- 

 coming preponderant. And, indeed, the observations show in 

 a way not to be doubted, that the general system of dips, of 

 variations, and of magnetic intensities, is very sensibly modi- 

 fied, sometimes in a sudden and irregular manner, by the neigh- 

 bourhood of great chains of mountains. This appears to be 

 confirmed by the singular inflexion which the magnetic equa- 

 tor experiences near the numerous archipelagos of the South 

 Sea. It is well known, indeed, that the islands with which this 

 sea is scattered, are only the summits of very high mountaitis 

 which elevate themselves quite to a peak from the midst of an 

 ocean where we can find no bottom. If the madrepores, of 

 \vhich they appear to be composed, only form a thin layer, and 

 if, as has been supposed by some very able naturalists, the re- 

 mainder of their mass has been produced by the action of sub- 

 terraneous fires, the system of these islands would form the most 

 extensive volcanic chain that there is on the svnface of the globe. 

 Then, all the irregularities produced by this system in the ge- 

 neral laws of terrestrial magnetism, would have nothing in them 

 but what is simple, and conformable to what is observed in vol- 

 canic countries*. For the action of subterraneous fires must 



necessarily 



* If tlie great niai;i.ct of the ewtli be like tnanv loarlstoiies that are to be 

 niet'with, aiid have more th;iii tw(j jiule;-, a* tiie iiii^ciiious author supposes, 



these 



