MAGNETIC VARIATION. 145 



tude of 68, in the meridian of Jakutsk 129 50' E. long., 

 forming at this point the outer shell of the eastern group of 

 oval concentric lines of variation, to which we have frequent- 

 ly referred, again sinking in the direction of Ochotsk in 143 

 10' E. long., intersecting the arc of the Kurile Islands, and 

 penetrating into the southern part of the Japanese Sea. All 

 the curves of from 5 to 15 eastern variation which occupy 

 the space between the lines of no variation in Western arid 

 Eastern Asia have their concavities turned northward. The 

 maximum of their curvature falls, according to Erman, in 

 80 E. long., and almost in one meridian between Omsk and 

 Tomsk, and are therefore not very different from the merid- 

 ian of the southern extremity of the peninsula of Hindos- 

 tan. The axis major of the closed oval group extends 28 

 of latitude as far as Corea. 



A similar configuration, although on a still larger scale, 

 is exhibited in the Pacific. The closed curves here form 

 an oval between 20 N. lat. and 42 S. lat. The axis ma- 

 jor lies in 130 W. long. That which most especially dis- 

 tinguishes this singular group (the greater portion of which 

 belongs to the southern hemisphere, and exclusively to the 

 sea) from the continent of Eastern Asia is, as has been al- 

 ready observed, the relative succession in the value of the 

 curves of variation. In the former the eastern variation di- 

 minishes, while in the latter the western variation increases 

 the farther we penetrate into the interior of the oval. The 

 variation in the interior of this closed group in the southern 

 hemisphere amounts, however, as far as we know, only to 

 from 8 .to 5. Is it likely that there is a ring of southern 

 variation within the oval, or that we should again meet with 

 western variation farther to the interior of this closed line of 

 no variation ? 



Curves of no variation, like all magnetic lines, have their 

 own history, which, however, does not as yet, unfortunately, 

 date further back than two centuries. Scattered notices may 

 indeed be met with as early even as in the 14th and 15th 

 centuries ; and here, again, Hansteen has the great merit of 

 having collected and carefully compared together all the va- 

 rious data. It would appear that the northern magnetic pole 

 is moving from west to east, and the southern magnetic pole 

 from east to west ; accurate observations show us, however, 

 that the different parts of the isogonic curves are progressing 

 very irregularly, and that where they were parallel they are 

 losing their parallelism ; and, lastly, that the domain of the 

 VOL. V. G 



