VOL. XI 

 1891 



"'] PROCEEDINGS OF THE NATIONAL MUSEUM. 21 



The Guianas present more fauiial similarity to the Amazons than to 

 the Orinoco, notwithstanding the fact that a direct connection between 

 the Rio Negro and Cassiquiare exists. The anomaly may be explained 

 by the comparative state of knowledge of the Orinoco and the Guianas, 

 the former having received but little attention from explorers, while the 

 Guianas — especially British Guiana — have been pretty well searched by 

 many naturalists. 



The genera of the Amazons (Lists vir, viii, ix, x) are sufficient in 

 number to warrant the separation of the Amazons, exclusive of the 

 high mountain sources, as a distinct province, the Amazonian province. 

 This province ought probably to include the Orinoco. As a conven- 

 ience the genera are separated into four divisions, but many of Lists 

 VIII to X will certainly be placed under the head of vii when the geograph- 

 ical limits become better known. Many of the genera enumerated under 

 VII are known from only two portions of the large system, some being 

 from the Amazon and Solimoens, others from the Solimoens and Mar- 

 anon, and others from the Amazon and Marahon. The last combina- 

 tions are again to be explained by our comparative lack of knowledge 

 of the Solimoens fauna. There are also genera (as Mesonauta) found 

 in the Amazon and the Guaporc under conditions which differ more than 

 those between the Amazon and the Maranon. For the present, then, 

 the whole of the Amazon basin may be considered as one province. 

 The Amazon fauna presents many similarities to the Guiana fauna. 



Lists XI to XIII show in a striking manner the paucity of peculiar 

 generic types in the San Francisco, Guianas, and Magdalena regions and 

 their entire absence in the Orinoco region. The comparatively large 

 number of genera peculiar to the Guianas is doubtless due to the large 

 number of isolated river systems which are yet too closely united to 

 warrant a separation into distinct provinces. The absence of peculiar 

 generic f^'P^s in the Orinoco is probably due to our meager knowledge 

 of that large river and to its direct connection with the Amazons. 



The Rio Magdalena, considering its isolation and the fact that it lies 

 entirely to the west of one of the highest northern Andean ridges, has 

 remarkably few generic types peculiar to itself as well as a strikingly 

 large number of species found in other eastern rivers. If we compare 

 this with the paucity of identical types in the Pacific province and in 

 the eastern provinces we have before us a self-evident proof that, within 

 a certain limit, bodies of salt water i^resent a much weaker barrier to 

 the distribution of fresh-water fishes of South America than even a 

 narrow and comjiaratively low mountain chain such as separates the 

 Oauca from the Pacific province. 



The Rio San Francisco has but two peculiar genera which are very 

 closely related to genera of wide distribution. This can not be attrib- 

 uted to lack of knowledge, for, through the labors of Reinhardt, Liitken, 

 and others, this river fauna has been as well made known as that of any 

 other region. As will be seen from List vi, this river has several genera 



