563 



terior course is no longer known, continues to 

 be two thirds of the breadth of the Rio Negro 

 near San Carlos. This opinion appears to me 

 less probable ; but I relate what I have collected, 

 and affirm nothing positively. I know by the 

 numerous measurements which I have taken, 

 how easily we may be deceived in the dimensions 

 of the beds of rivers. In fact, rivers appear 

 more or less broad according as they are sur- 

 rounded by mountains or plains, free from islets, 

 or full of shoals, swelled by violent rains, or 

 bereft of their waters by long drought. We 

 may recollect also, that the course of the Ganges 

 is unknown to the north of Gangootra ; and that, 

 on account of it's little breadth, this point is be- 

 lieved to be very near it's source. 



In the rocky dike that crosses the Oroonoko, 

 forming the Raudal of the Guahariboes, Spanish 

 soldiers pretend to have found the fine kind of 

 saussurite (amazon stone), of which we have 

 spoken above. This tradition however is very 

 uncertain ; and the Indians, whom I interro- 

 gated on the subject, assured me, that the green 

 stones, called piedras de Macagua * at Esme- 

 ralda, were purchased from the Guaicas and 



* See above, p. 383. The etymology of this name, which is 

 unknown to me, might lead to the knowledge of the spot, 

 where these stones are found. I have sought in vain the name 

 of Macagua among the numerous tributary streams of the 

 Tacutu, the Mahu, the Rupunury, and the Rio Trombetas. 



2o2 



