72 ANNALS NEW YORK ACADEMY OF SCIENCES 



■duced some dissimilar changes. Many of these changes are of such a 

 nature that the species are adapted to live only in certain kinds of en- 

 vironments. The cichlid fishes usually live in swamps, lagoons or lakes, 

 and seldom in rapidly flowing water. In the open channels, the chara- 

 cins would eat them. The Paraguay and the Amazon have many swamps 

 and many cichlid fishes. Alto Eio Parana has few swamps and few 

 •cichlids. 



The marine immigrants which have entered the rivers and become 

 permanent dwellers of the same have also increased the percentage of 

 similarity between certain rivers more than between others. Typical 

 •examples of such fishes are as follows : 



1. The fresh-water skates, Potamotrygon, two species of which are 

 found both in the Paraguay and the Amazon. We have no evidence that 

 these two species separately left the ocean and became dwellers of these 

 rivers. Other species of the same genus are found only in the Amazon, 

 in the La Plata, the Guiana and the Orinoco. Is it not possible that some 

 of these species are the results of changes invoked in the marine ancestor 

 "by their new environments? 



2. Two species of Poeciliidse, Rivulus and Girardinus, could easily have 

 gone along the coast from the La Plata to the Amazon. They may also 

 have gone overland. 



3. Stoleophorus olidus. 



4. Two species of Sciasnidse belonging to the genus Pachyurus. 



5. Several catfishes, and even a few of the characinids and cichlids, 

 might have migrated along the coast, but it is out of the question to as- 

 sume that all of the identical forms in Eio Paraguay and Eio Amazonas 

 did so because they are not found in coastwise streams, Eio Uruguay and 

 Uio Parana. 



There is no doubt that the marine immigrants have played an im- 

 portant part in the production of a greater similarity between the fresh- 

 water fishes of certain river basins than others. I found only two marine 

 species of fish in Eio Colorado of Patagonia and at least one hundred are 

 found in the lower Amazon. Even the sawfish (Pristis) is sometimes 

 killed as far up the Amazon as Santerem (476 miles). 



The volume of water always bears a relation to the number and size of 

 marine immigrants, and this is especially true when there are many 

 islands, many channels, plenty of food, tidal effects and much brackish 

 water. 



It is evident, then, that the small rapid coastwise streams of eastern 

 Brazil, Bio Magdalena and Patagonia should have fewer species than 



