of the Surface of the Earth. 113 



valley. It would seem as if, in its case, a flood of water bad 

 directed its violence, first towards the north-east, then to- 

 wards the north west, and then again towards the north-east. 

 The parallelism of the coasts to the north of 10^ of south lat., 

 the salient and re-entering angles, the convexity of Brazil 

 opposite the Gulf of Guinea, and the convexity of Africa in 

 the same latitude with the Gulf of the Antilles, are all in 

 favour of this apparently bold idea.* In the Atlantic valley, 

 as is the case almost everywhere with the configuration of 

 great masses of land, coasts which are indented and abound 

 in islands are placed opposite to those which are of a imi- 

 form character. I have long since directed attention to the 

 geographical interest of a comparison of the west coasts of 

 Aft'ica and South America in the tropical zone. The gulf- 

 like indentation of the African shore at Fernando Po (north 

 lat. 4-^^), is repeated on the shore of the South Sea, in south 

 lat. 18^°, near Arica, where (between the Valle de Arica and 

 the Morro de Juan Diaz) the Peruvian coast suddenly alters 

 its direction from a south and north one to a north-western 

 one. This change of direction extends, in an equal degi*ee, 

 to the two parallel ranges of the chain of the Andes : not 

 only in the littoral t high land, but also in the eastern, which 

 was the earliest seat of human cultivation in South America, 

 and where the alpine Lake of Titicaca is bounded by the 

 colossal mountains of Sorata and Illimani. Further to the 

 south, from Valdivia and Chiloe (south lat. 40^ to 42°), through 

 the Archipelago de los Chonos, as far as Tierra del Fuego, we 

 find again repeated the peculiar /?orf/-arrangement (the 

 complication of narrow and long indentations) which, in the 

 northei'n hemisphere, is characteristic of the west coasts of 

 Norway and Scotland. 



* Humboldt in the Journal de Physique, t. liii., 1799, p. 33, and Rel. 

 Hist., t. ii., p. 19 ; t. iii., pp. 189 and 198. 



t Humboldt in Fo(!gcndorff's Annalcn dcr Physih, vol. xl., p. 171. As 

 to the remarkable Fiords of the south-east extremity of America, see 

 Darwin's Journal (Narrative of the Voyages of the Adventure and Beagle, 

 vol. iii.), 1839, p. 2QQ. The parallelism of the two mountain-chains con- 

 tinues from 5" south lat. to 5" north lat. The change of the direction of 

 the coast near Arica appears to be the consequence of the altered strike of 

 the fissure through which the Cordillera de los Andes has been elevated. 



VOL. XXXIX. NO. LXXVII. — JULY 1845. H 



