334 PATAGONIAN EXPEDITIONS: ZOOLOGY. 



from the San Francisco westward and not eastward. This applies with 

 equal force to all other San Franciscan genera found in the Amazon and 

 not in the coastwise streams.^ 



How is it with the genera Psendoplaty stoma, RJiamdia, Rhamdella, 

 Pimelodella, Pime/odiis, Astyanax and all the other genera of the San 

 Francisco that have a wide distribution?^ These probably belong to the 

 original fauna, or had their origin in the San Francisco, or rather Brazilian 

 plateau and have become distributed from it. Guiana has of course an 

 equal claim' on them ; and there is no evidence that they have become 

 distributed from the San Francisco rather than from Guiana. 



There remain only the genera peculiar to the region, and, inasmuch as 

 they have not become distributed they do not enter the question. 



There is some evidence that part of the San Francisco fauna has been 

 derived from the Amazon basin. This evidence consists of the fact that 

 the San Francisco, which is nearer the Amazon basin, contains a larger 

 number of Amazonian genera than the eastern streams, and in part, of the 

 fact that the peculiarly lowland forms of the Amazon have not reached this 

 province, and in part, of the presumption in favor of the Amazon because 

 it has unquestionably acted as the center of distribution of many lowland 

 forms that have reached the La Plata basin without entering the province 

 under consideration.- 



It is very probable that the San Francisco has contributed to the fauna 

 to the south of it (see under the La Plata, etc.). 



9. The Coastal Pyomnce. 



The fauna of the second terrace and coastal plain from Bahia to Iguape, 

 whose rivers were described in a preceding section, shows a proportion 

 of species peculiar to the region about equal to the proportion of peculiar 

 species in the Pacific province, not considering marine derivatives and 

 Poeciliidae. There are eight peculiar genera, as compared with three in the 

 Pacific province. This area contains a very much larger percentage of 

 species that are also found in the Amazon than does the Magdalena. 



Fifty genera present in the Amazon and La Plata basins are absent from 

 the large streams between Bahia and Rio de Janeiro ; over thirty of these 

 are also absent from the San Francisco. They demonstrate their migra- 



' See in the Guiana list those marked with an S. 



^ Mr. Haseman reports an open road for the passage of fishes from the Tocantins. 



