497 



the savannahs of Caraccas, Barcelona, and Co- 

 mana begin. Geographers, who are fond of 

 copying, and of stereotyping, for ages, the chains 

 of mountains and the branches of rivers which 

 the caprice of the draftsman has traced on 

 some well-known maps, never cease to figure, 

 between the meridians of Caraccas and Cu- 

 mana, two Cordilleras stretching from north to 

 south, as far as 81° of latitude ; to which they 

 give the names of Cerros de Alta Gracia, and 

 del Bergantin *; thus rendering a territory of 



* See all the French, English, and German maps pub- 

 lished before the Map of Columbia, by M. Brae* (1823) , for 

 which a part of the materials were employed which I had 

 collected on the extent and direction of the chains of moun- 

 tains. The source of the error which we find in Nicolosio, 

 Sanson (1669), and De l'lsle (1700), must be attributed to 

 the practice of the first geographers of America, of enlarg- 

 ing beyond measure, the breadth of the Andes of Peru and 

 New Grenada, and placing them so far towards the east, 

 that Quito is sometimes found on the meridian of Cumana 

 (Vol. v, p: 053). In this manner, the steppes of Vene- 

 zuela were covered with mountains that linked the groupe of 

 the Parime with the chains of the shore of Caraccas. De 

 l'lsle places the falley of Sayma near the range of moun- 

 tains which Sanson had marked as going from north to 

 south, from Barcelona to the Oroonoko j this proves that he 

 had some confused idea of the mountains of Caripe, inha- 

 bited by the Chaymas Indians. D'Anville, according to 

 systematic ideas on the origin of rivers, figures a ridge be- 

 tween the sources of the Unare, the Guarapiche, the Pao, 

 and the Manapire (Vol. iv, p. 301). This is the pattern 



VOL. VI. 2 L 



