TERRESTRIAL MAGNETISM. 147 



ellipse, like the Atlantic of 0, have a direction from 

 south to north, and between the Caspian and Lapland, 

 from SSE. to NNW. The third system, that of the 

 Pacific Ocean, which has been least examined, is less 

 than the others, and is almost entirely confined to the 

 south of the geographic equator : it forms a closed oval 

 of concentric lines of successively decreasing east de- 

 clination. On the African continent, inferring from 

 what we know of the declination on its coasts, ( 188 ) 

 there are only lines having a west declination of from 

 6 to 2 9; the Atlantic line of no declination quitted 

 the south point of Africa (Cape of Grood Hope), ac- 

 cording to Purchas, as early as 1605, moving from east 

 to west. As far as we are acquainted with the facts, 

 there is not in the interior of central Africa an oval 

 group of concentric declination lines similar to that in 

 the Pacific. 



The Atlantic part of the American curve of no de- 

 clination has been accurately determined and laid down 

 in an excellent work by Colonel Sabine, "On the De- 

 clination Lines over the Atlantic Ocean, between the 

 parallels of 60 north and 60 south latitude, for the 

 year 1840," based on 1480 observations either made in 

 that year, or reduced to it by careful allowance for 

 secular change. Tracing this line of no declination 

 from its most southern known point (in lat. 70 S., 

 where it was found in about 19 W. long.), ( 189 ) it runs 

 NNW. to 3 west of Cook's Sandwich Land, and to 9| 

 east of South Georgia ; it then approaches the Brazilian 

 coast, which it enters at Cape Frio, two degrees east of 

 Rio Janeiro ; cootinues within the South American 

 continent to S. lat. 36', and quitting the continent 



L 2 



