OEOGBAPHICAL DOUBTS. 339 



reflex verdure seems to hav< the same vivid hue as that 

 which clothes the real vegetation. The surface of the fluid 

 is homogeneous, smooth, and destitute of that mixture of 

 suspended sand and decomposed organic matter, which 

 roughens and streaks the surface of less limpid rivers. 



On quitting the Orinoco, several small rapids must be 

 passed, but without any appearance of danger. Amid these 

 raudalitos, according to the opinion of the missionaries, the 

 Rio Atabapo falls into the Orinoco. I am however disposed 

 to think that the Atabapo falls into the Guaviare. The Rio 

 Guaviare, which is much wider than the Atabapo, has white 

 waters, and in the aspect of its banks, its fishing-birds, its 

 fish, and the great crocodiles which live in it, resembles the 

 Orinoco much more than that part of the Atabapo which 

 comes from the Esmeralda. When a river springs from the 

 junction of two other rivers, nearly alike in size, it is difficult 

 to judge which of the two confluent streams must be re- 

 garded as its source. The Indians of San Fernando affirm 

 that the Orinoco rises from two rivers, the Guaviare and 

 the Rio Paragua. They give this latter name to the Upper 

 Orinoco, from San Fernando and Santa Barbara to beyond 

 the Esmeralda, and they say that the Cassiquiare is not an 

 arm of the Orinoco, but of the Rio Paragua. It matters but 

 little whether or not the name of Orinoco be given to the 

 Rio Paragua, provided we trace the course of these rivera 

 as it is in nature, and do not separate by a chain of moun- 

 tains, (as was done previously to my travels,) rivers that 

 communicate together, and form one system. When we 

 would give the name of a large river to one of the two 

 branches by which it is formed, it should be applied to that 

 branch which furnishes most water. Now, at the two 

 seasons of the year when I saw the Guaviare and the Upper 

 Orinoco or Rio Paragua (between the Esmeralda and San 

 Fernando), it appeared to me that the latter was not 

 so large as the Guaviare. Similar doubts have been 

 entertained by geographers respecting the junction of the 

 Upper Mississippi with the Missouri and the Ohio, the 

 junction of the Marafion with the Guallaga and the Uca- 

 yale, and the junction of the Indus with the Chunab 

 (Hydaspes of (Cashmere) and the Gurra, or Sutlej.* To 



* The Hydaspes is properly a tributary stream of the Chunab o* 



E 2 



