TRUE VOLCANOES. 293 



With regard to the geological causal connection of the agree- 

 ment, which is so often manifested between the outlines of 



the volcanic chain of the Andes of Chili, Peru, Bolivia, Quito, and 

 New Granada, be connected with the mountain chain of the Isthmus 

 of Panama, and in this way with that of Veragua and the series of vol- 

 canoes of Costa Pdca and Central America in general. In my maps 

 of 1816, 1827, and 1831, the mountain-systems of which have been 

 made more generally known by Bru in Joaquin Acosta's fine map of 

 New Granada (1847) and in other maps, I have shown how the chain 

 of the Andes undergoes a triple division under the northern parallel 

 of 2 10'; the western Cordillera running between the valley of the 

 Rio Cauca and the Rio Atrato; the middle one between the Cauca and 

 the Rio Magdalena; and the eastern one between the valley of the 

 Magdalena and the Llanos (plains) which are watered by the affluents 

 of the Maranon and Orinoco. I have been able to indicate the special 

 direction of these three Cordilleras from a great number of points 

 which fall in the series of astronomical local determinations, of which 

 I obtained 152 in South America alone by culminations of stars. 



To the east of the Rio I^gua, and to the west of Cazeres, Rolda- 

 nilla, Toro, and Anserma, near Cartago, the western Cordillera runs 

 S.S.W. N.N.E., as far as the Salto de San Antonio in the Rio Cauca 

 (lat. 5 14') which lies to the south-west of the Vega de Supia. Thence, 

 as far as the Alto del Viento (Cordillera de Abibe, or Avidi, lat. 7 12') 

 9600 feet in height, the chain increases considerably in elevation and 

 bulk, and amalgamates, in the province of Antioquia, with the inter- 

 mediate or Central Cordillera. Further to the north, towards the 

 sources of the Rios Lucio and Guacuba, the chain ceases, dividing into 

 ranges of hills. The Cordillera occidental, which is scarcely 32 miles 

 from the coast of the Pacific near the mouth of the Dagua in the 

 Bnhia de San Buenaventura (lat. 3 50*) is twice this distance in the 

 parallel of Quibdo in the Choco (lat. 5 48'). This observation is of 

 some importance, because we must not confound with the western 

 chain of the Andes, the country with high hills, and the range of hills, 

 which in this province, so rich in gold dust, runs from south to north 

 from Xovita and Tado along the right bank of the Rio San Juan and 

 the left bank of the great Rio Atrato. It is this inconsiderable series of 

 hills that is intersected in the Quebrada de la Raspadura, by the 

 canal of Raspadura (Canal des Monches), which unites two rivers (the 

 Rio San Juan or Noanama and the Rio Quibdo, a tributary of the 

 Atrato) and by their means two oceans (Humboldt, Essai Politique,^,. i, 

 p. 235); it was this also which was seen in the instructive expedition of 

 Captain Kellet between the Bahia de Cupica (lat. 6 42*) long and fruit- 

 lessly extolled by me, and the sources of the Napipi, which falls 

 into the Atrato. (See Humboldt, Op. cit. t. i, p. 231 ; and Robert 

 Fitzroy, Congiderations on the Great Isthmus of Central America, in 

 the Journal of the Royal Gcogr. Soc. vol. xx, 1851, pp. 178, 180, and 

 186). 



The middle chain of the Andes (Cordillera Central), constantly tha 

 highest, reaching within the limit of perpetual snow, and in ita entire 



