HASEMAN, GEOGRAPHICAL DISTRIB UTION IN SO UTH AMERICA 39 



feet high. Similar conditions are said to exist between the headwaters of' 

 Eios Paraguay, Tapajos and Xingu, all of which rise on the Piano Alto 

 and are not connected. 



Stream piracy on a large scale has taken place between Eio Sao Fran- 

 cisco and Eio Tocantins and between Eio Negro and Eio Orinoco. 



Eio Orinoco is robbing Eio Negro of one of its previous affluents, the 

 Canal de Cassiquiare. This is due to the fact that there is a long gentle 

 slope of the highlands toward the Amazon Valley, while the northern and 

 eastern slopes are more abrupt. On account of these conditions, Eio 

 Orinoco is still cutting rapidly and deeply through a so-called sierra just 

 above its connection with the Canal de Cassiquiare. The Eio Negro on 

 the other hand is not cutting as rapidly in this region as Eio Orinoco. 



The Canal de Cassiquiare is a long, narrow, lakelike mass of water with 

 high sand hills along and near either bank. The Eio Negro end of this 

 "canal" is being gradually stopped up by secondary deposits of sand and 

 plants, and consequently as soon as Eio Orinoco cuts more deeply into the 

 hard basal highland rock, it will drain the Canal de Cassiquiare and 

 appropriate it as an affluent. 



Lakes and swamps similar to the Canal de Cassiquiare are quite com- 

 mon in and on all of the Piano Alto. Sete Lagoas, at the headwaters of 

 Eio Paraguay, and Lagoa Feia, near Paraeatu, in southeastern Goyaz, are 

 examples of such lakes or swamps; but inasmuch as no marine deposits 

 are known to exist in any of these regions, it appears that all such lakes 

 are entirely due to unequal erosion and changes in the courses of the high- 

 land streams. Such is evidently the case of Canal de Cassiquiare, for it is 

 almost surrounded by the highland sandstone, and in the regions of the 

 upper waterfalls both Eio Orinoco and Eio Negro flow over the hard 

 ancient basal highland rock. The rocks in the Eio Negro Falls are essen- 

 tially the same as those of the Madeira Falls. 



Similar conditions exist between the headwaters of Eio Tocantins and 

 Eio Sao Francisco in the Jalapao portion of northern Goyaz. Wells 

 (1875) and Froh (1907) reported a connection between these rivers; but 

 my discovery, which was entirely independent, was based on the presence 

 of certain Amazonian fishes not previously reported from the Sao Fran- 

 cisco Eiver. 20 



Plate XIV shows that an affluent of Eio Tocantins and one of Eio Sao 

 Francisco rise in the Brejo de Varedao, which is surrounded on all sides 



20 It may be of some interest to note that I do not now consider that the presence of 

 these fishes is evidence for this connection, because I later found the same fishes in other 

 rivers where no connections exist. My erroneous assumption, however, which is a com- 

 mon one among static zoogeographers, led me to the discovery of several facts. 



