40 



as well as the hornblende slates *, which I sa\v 

 near Angostura, on the banks of the Lower 

 Oroonoko, belong to the mountains of Pacaraimo 

 and of Pari me, stretching from West to East^ 

 in the interior of the continent, and not in a 

 direction parallel with the coast, between the 

 mouths of the River of Amazons and the Oroo- 

 noko. But notwithstanding we find no chain 

 of mountains at the North-East extremity of 

 Terra Firma, that has the same direction as the 

 Archipelago of the Smaller West India islands, 

 it does not follow from this circumstance alone, 

 that the volcanic mountains of the Archipelago 

 may not have belonged originally to the conti- 

 nent, to the littoral chain of Caraccas and 

 Cumana 



* Hornklendschiefer, amphibolites schistoides of Brongniart. 



+ From the cataracts of Atures toward the Rio Esquibo. 

 This chain of Pacaraimo divides the waters of the Carony 

 from those of the Rio Parime, or Rio de Aguas Blancas. 

 See my Analyse de V Atlas geograph. PI. xvi. 



+ Among the numerdus examples, which the structure of 

 the Globe displays, we shall mention only the inflexion at 

 a right angle formed by the High Alps toward the maritime 

 Alps, in Europe; and the Belour-tagb, which joins trans- 

 versely the Mouzitagh with Himalaya, in Asia. Amid the 

 prejudices, which impede the progress of mineralogical geo- 

 graphy, we may reckon, 1st, The supposition of a perfect 

 uniformity of direction in the chains of mountains ; 2d, The 

 hypothesis of a continuity in all the chains j 3d, The suppo- 

 sition, that the highest summits determine the direction of a 



